Global gene expression changes in Candidatus Liberibacter asiaticus during the transmission in distinct hosts between plant and insect

Summary

Huanglongbing (HLB) or citrus greening disease is a destructive disease of citrus worldwide, which is associated with Candidatus Liberibacter asiaticus. This phloem-limited fastidious pathogen is transmitted by the Asian citrus psyllid, Diaphorina citri, and appears to be an intracellular pathogen that maintains an intimate association with the psyllid or the plant throughout its life cycle. The molecular basis of the interaction of this pathogen with its hosts is not well understood. We hypothesized that, during infection, Ca. L. asiaticus differentially expresses the genes critical for its survival and/or pathogenicity in either host. To test this hypothesis, quantitative reverse transcription-polymerase chain reaction was performed to compare the gene expression of Ca. L. asiaticus in planta and in psyllid. Overall, 381 genes were analysed for their gene expression in planta and in psyllid. Among them, 182 genes were up-regulated in planta compared with in psyllid (P < 0.05), 16 genes were up-regulated in psyllid (P < 0.05) and 183 genes showed no statistically significant difference (P ≥ 0.05) in expression between in planta and in psyllid. Our study indicates that the expression of the Ca. L. asiaticus genes involved in transcriptional regulation, transport system, secretion system, flagella assembly, metabolic pathway and stress resistance are changed significantly in a host-specific manner to adapt to the distinct environments of plant and insect. To our knowledge, this is the first large-scale study to evaluate the differential expression of Ca. L. asiaticus genes in a plant host and its insect vector.

Introduction

In order to adapt to diverse environments, bacteria alter their gene regulation. Bacteria are known to sense and respond to changes in nutrients, pH, temperature, oxygen tension, redox potential and osmolality for optimal growth and survival in different environmental niches (Mekalanos, 1992). Bacterial gene expression is altered by the activation of specific genes whose products assist in survival and by the deactivation of those whose products are not necessary in a particular environment (Chowdhury et al., 1996). Plant pathogenic bacteria are known to up-regulate genes which are implicated in bacterial adaptation to the host environment during infection. As reported by Okinaka et al. (2002), Erwinia chrysanthemi genes encoding products involved in anaerobiosis, iron uptake, transporters, chemotaxis and stress responses to reactive oxygen species (ROS) and heat are induced in planta. It has been suggested that virulence factors are expressed at different stages of the infection process dictated by the changing microenvironment of the host (Chowdhury et al., 1996).

Interestingly, insect-transmitted bacterial plant pathogens switch hosts belonging to two distinct kingdoms, namely plants (Plantae) and insects (Animalia). These different environments seem to have a dramatic effect on the gene regulation of bacterial plant pathogens. Oshima et al. (2011) have carried out pioneering work comparing the gene expression of Candidatus Phytoplasma asteris OY-M in garland chrysanthemum (Chrysanthemum coronarium) and its leafhopper vector, Macrosteles striifrons. Approximately 33% of Phytoplasma genes showed differential expression on switching between plant and insect hosts, suggesting that Phytoplasma regulates genes encoding transporters, secreted proteins and metabolic enzymes in a host-specific manner. The elucidation of the mechanism underlying the adaptation of a pathogen to different hosts may contribute to the development of novel methods for disease control tactics.

Among the insect-transmitted bacterial plant pathogens, Candidatus Liberibacter asiaticus is associated with citrus huanglongbing (HLB), which is the most devastating disease of citrus worldwide. This disease is widespread in the citrus-growing regions of Asia, Africa and the Americas, resulting in a substantial reduction in the lifespan of infected trees and fruit production (Bové, 2006). Candidatus L. asiaticus is capable of infecting all known commercial varieties of citrus and several close relatives, although symptom expression varies depending on the host genotype (Folimonova et al., 2009; Sagaram et al., 2009; Tyler et al., 2009). Characteristic symptoms associated with HLB include blotchy mottle and the yellowing and thickening of leaves, together with enlarged midribs and lateral veins (Bové, 2006). The infected leaves grow upright and, eventually, defoliation followed by dieback may occur. HLB also affects fruit quality and appearance, and interferes with sucrose transport, resulting in heavy starch accumulation, phloem damage and eventual phloem collapse (Etxeberria et al., 2009; Folimonova and Achor, 2010; Kim et al., 2009).

Candidatus L. asiaticus is transmitted by the Asian citrus psyllid Diaphorina citri. The insect transmits the bacterium via salivary secretion in a persistent manner (Hung et al., 2004; Xu et al., 1988). Diaphorina citri ingests the bacteria whilst feeding on the phloem of infected plants, and the bacteria circulate inside its vector, passing from the midgut into the haemolymph, and then migrating to the salivary glands, from which they are inoculated into the plant phloem with saliva during subsequent probing (Ammar et al., 2011). Although the psyllid can acquire the pathogen during nymphal and adult stages (Inoue et al., 2009; Xu et al., 1988), the adults are more successful in transmitting the bacteria when they acquire them as a nymph (Pelz-Stelinski et al., 2010). Once infected, D. citri is capable of inoculating the host plant with Ca. L. asiaticus throughout its life (Hung et al., 2004), although the level of pathogen transmission decreases over time (Pelz-Stelinski et al., 2010).

Candidatus L. asiaticus maintains an intimate association with the psyllid or plant throughout its life cycle. Being a fastidious intracellular pathogen, the chromosome of Ca. L. asiaticus has undergone significant reduction compared with other members of the Rhiozobiaceae family. Candidatus L. asiaticus lacks the common virulence determinants, including the type III (T3SS) and type II (T2SS) secretion systems (Duan et al., 2009). Therefore, information regarding the virulence determinants of this pathogen and the mechanism(s) of interaction of Ca. L. asiaticus with its hosts remains limited. A clear understanding of the molecular basis of the interaction between Ca. L. asiaticus and its hosts is essential to devise control measures against HLB.

In this study, we compared the gene expression of Ca. L. asiaticus in its plant host, sweet orange (Citrus sinensis) ‘Valencia’, and its insect host, the psyllid D. citri. The implication of the gene expression pattern to the pathogen adaptation of Ca. L. asiaticus to its distinct hosts is discussed.

Results and Discussion

Global gene expression profiling of Ca. L. asiaticus in planta and in psyllid

To compare the gene expression of Ca. L. asiaticus in its plant host, sweet orange (C. sinensis) ‘Valencia’, and its insect host, psyllid D. citri, quantitative reverse transcription-polymerase chain reaction (QRT-PCR) was conducted. QRT-PCR is sensitive and reliable for the measurement of gene expression (Erickson et al., 2009). It is possible to measure the gene expression of Ca. L. asiaticus using QRT-PCR because of its small genome size of 1.23 Mb. It contains 1136 predicted open reading frames (ORFs), approximately 74% of which have homologues with known and putative functions, whereas the other 26% represent hypothetical proteins (Duan et al., 2009). To further narrow down the genes being tested, we neglected most housekeeping genes, and focused on putative virulence genes, such as genes with domains and motifs found in known virulence factors, and genes involved in transcriptional regulation, the metabolic pathway, secretion system, transportation, motility and signal transduction. In total, 523 genes were selected (Table S1, see Supporting Information) for QRT-PCR analysis, and gene-specific primers were designed (Table S2, see Supporting Information). Among them, 43 genes were discarded because of lack of amplification and 99 as a result of nonspecific amplification (data not shown). We were able to optimize the primers and to perform specific amplification for 381 genes, that were subsequently analysed for their gene expression in planta and in psyllid.

It should be noted that, unlike the Ca. L. asiaticus-infected plant samples, which were collected according to visible symptoms, no visible difference was observed between infected and noninfected psyllid samples. Thus, two methods were used to collect psyllid RNA samples subjected to QRT-PCR in this work: (i) 15–20 psyllids were pooled before RNA extraction; (ii) RNA was extracted from a single psyllid and only Ca. L. asiaticus-infected psyllids (Fig. S1, see Supporting Information) were used for further analysis. To validate the gene expression profiles of Ca. L. asiaticus, the expression of 21 randomly selected genes was compared by QRT-PCR using the two differently isolated psyllid RNA samples. The results showed a high degree of concordance (correlation coefficient R² = 0.73) between the data generated by the two methods (Fig. 1). This result indicates that both methods could reliably detect the gene expression of Ca. L. asiaticus in psyllid.

Figure 1

Correlation of gene expression profile of Candidatus Liberibacter asiaticus in planta (Citrus sinensis) and in psyllid (Diaphorina citri) using two methods to collect psyllid RNA samples. The relative expression values (in planta versus in psyllid, log₂ ratio) of 21 genes were plotted. R², correlation coefficient.

To identify genes with statistically significant changes in expression, only those with a P value for a t-test of less than 0.05 are discussed further in this work. Among the genes selected for the expression test, 198 showed a statistically significant expression change in the host plant or insect. Specifically, 182 genes were up-regulated significantly in planta compared with in psyllid; 16 genes were up-regulated significantly in psyllid. Using the Clusters of Orthologous Groups (COG) Database, the 198 genes were categorized into 20 diverse functional groups plus one group of genes not classified in the COG Database (Fig. 2). In addition to the hypothetical genes, genes involved in cell motility constituted a major portion of the genes up-regulated in psyllid (Fig. 2). Large numbers of genes involved in transcriptional regulation, metabolism, transport and cell membrane biosynthesis were up-regulated in planta (Fig. 2).

Figure 2

Functional classification of differentially expressed genes of Candidatus Liberibacter asiaticus. A total of 198 genes of Ca. L. asiaticus showed significant expression changes (P < 0.05) in planta (Citrus sinensis) and in psyllid (Diaphorina citri) hosts. Using the Clusters of Orthologous Groups (COG) Database, the 198 genes were classified into 20 functional groups plus one group of genes that were not classified in COG. Red represents genes up-regulated in planta compared with in psyllid; blue represents genes up-regulated in psyllid compared with in planta.

Transcriptional factors

Consistent with the large numbers of genes of Ca. L. asiaticus up-regulated in the host plant, 11 genes encoding (putative) transcriptional regulators were up-regulated in planta (Table 1). Sigma factors are transcriptional regulators that interact with the core RNA polymerase and direct the initiation of transcription at cognate promoter sites. Two sigma factors, RpoD and RpoH, were annotated in the genome of Ca. L. asiaticus. In this work, we found that the expression of rpoH (CLIBASIA_02490) was up-regulated significantly in the host plant. RpoH, also called alternative sigma factor 32, plays an important role in response to diverse environmental stresses, including heat shock, pH and oxidative stress. In addition, RpoH has also been reported to be involved in virulence and growth in the hosts of the human pathogens Brucella melitensis and Vibrio cholera (Delory et al., 2006; Slamti et al., 2007). The significant up-regulation of rpoH indicates that this sigma factor may be important in the gene regulation of Ca. L. asiaticus during its adaptation to environmental changes between insect and plant.

Table 1. Expression profile of genes encoding transcriptional factors of Candidatus Liberibacter asiaticus in planta compared with in psyllid

Locus tag	Gene nameb	Gene productc	Relative expressiond
CLIBASIA_00835a	NA	LysR family transcriptional regulator	2.23
CLIBASIA_00985a	NA	Two-component response regulator CpdR	1.61
CLIBASIA_01180	NA	MarR family transcriptional regulator	5.23
CLIBASIA_01510a	NA	CarD family transcriptional regulator	1.24
CLIBASIA_01805a	NA	Two-component response regulator. CheY-like receiver domain	1.08
CLIBASIA_02490a	rpoH	Alternative sigma factor RpoH (sigma factor 32)	1.37
CLIBASIA_02535a	rirA	Iron-responsive transcriptional regulator	1.04
CLIBASIA_02900	NA	Putative LuxR family transcriptional regulator	2.11
CLIBASIA_02905	NA	Putative LuxR family transcriptional regulator	2.75
CLIBASIA_02950	phoU	Putative phosphate transport system protein	3.49
CLIBASIA_03950a	ctrA	Two-component response regulator	1.74

• ^a Psyllid RNA samples were extracted from a single psyllid (see Experimental procedures).
• ^b NA, gene name was not assigned.
• ^c Information on gene product was based on the annotation of the genome of Ca. L. asiaticus strain psy62 and definition in the KEGG database.
• ^d Fold change (log₂ ratio) is the relative gene expression (in planta versus in psyllid) of Ca. L. asiaticus. Positive value indicates that gene was overexpressed in planta compared with in psyllid; negative value indicates that gene was overexpressed in psyllid compared with in planta.

The expression of two putative LuxR family transcriptional regulators (CLIBASIA_02900, CLIBASIA_02905) was up-regulated in the host plant (Table 1). Regulators of the LuxR family are known to regulate gene expression in the process of quorum sensing. Quorum sensing is an intraspecies or interspecies cell–cell communication system widely distributed in bacteria, and controls multiple bacterial behaviours, including symbiosis, motility, biofilm and virulence (Waters and Bassler, 2005). For example, mutation of the quorum sensing (las or rhl) in Pseudomonas aeruginosa PAO1 reduced rhamnolipid production and elastase activity, which are known virulence factors in this pathogen (Pearson et al., 1997). Quorum sensing is also required for virulence and insect transmission of Xylella fastidiosa (Chatterjee et al., 2008b), an insect-vectored pathogenic bacterium causing Pierce's disease on grape and variegated chlorosis on citrus (Chatterjee et al., 2008a). However, no gene(s) encoding the AHL synthase (luxI) homologue was observed in Ca. L. asiaticus (Duan et al., 2009). How Ca. L. asiaticus coordinates its gene expression using LuxR remains to be determined.

Another interesting result was the up-regulation of an iron response regulator-encoding gene, rirA (CLIBASIA_02535), in planta (Table 1). Iron is an essential micronutrient for almost all known organisms. Highly efficient iron acquisition systems have been evolved by bacteria to scavenge iron from living niches (Andrews et al., 2006). However, the overloading of iron is deleterious to bacteria, mainly because of the formation of hydroxyl radicals that strongly react with all kinds of biomolecules (Braun, 1997). RirA (rhizobial iron regulator) was initially identified in Rhizobium leguminosarum as a negative regulator of iron uptake (Todd et al., 2002). A similar function of RirA homologues was also found in Sinorhizobium meliloti and Agrobacterium tumefaciens (Chao et al., 2005; Ngok-Ngam et al., 2009). In this work, two genes encoding the ABC transport system involved in iron uptake and four genes involved in haem biosynthesis were dramatically up-regulated in planta (Tables 2 and 3). The up-regulation of rirA might be a protective measure for Ca. L. asiaticus to prevent the deleterious consequences caused by potential iron overload.

Table 2. Expression profile of genes encoding transport systems of Candidatus Liberibacter asiaticus in planta compared with in psyllid

Locus tag	Gene nameb	Gene productc	Relative expressiond
CLIBASIA_00090	NA	Putative ATP-binding component of ABC-type transport system involved in resistance to organic solvents	7.60
CLIBASIA_00095	NA	Putative substrate-binding protein of ABC-type transport system involved in resistance to organic solvents	2.43
CLIBASIA_00265	NA	Periplasmic binding protein of ABC transporter involved in general l-amino acid transport	−1.29
CLIBASIA_00275	aapM	General l-amino acid transport system permease	3.64
CLIBASIA_00540	NA	ABC transporter permease	2.52
CLIBASIA_00790	NA	Putative ATPase components of ABC transporters with duplicated ATPase domains	4.52
CLIBASIA_01135	proX	Glycine betaine/proline ABC transporter	1.84
CLIBASIA_01140	NA	Predicted Co/Zn/Cd cation transporters	3.88
CLIBASIA_02120	NA	Iron(ii)/manganese ABC transporter (homologue of SitA)	7.92
CLIBASIA_02135	NA	Iron(ii)/manganese ABC transporter (homologue of SitD)	5.72
CLIBASIA_02415	NA	ABC transporter nucleotide binding/ATPase protein. Sulphonate/nitrate/taurine transporter	2.41
CLIBASIA_02420	NA	ABC transporter permease. Sulphonate/nitrate/taurine transporter	7.24
CLIBASIA_02955	pstB	ABC transporter, nucleotide binding/ATPase protein. Phosphate transporter	7.28
CLIBASIA_02960	pstA	ABC transporter, permease. Phosphate transporter	2.23
CLIBASIA_02965	pstC	ABC transporter, membrane spanning protein. Phosphate transporter	3.28
CLIBASIA_02970	pstS	Putative phosphate-binding periplasmic protein. Phosphate transporter	1.28
CLIBASIA_03155	lptB	ABC transporter nucleotide binding/ATPase protein. Lipopolysaccharide ABC transporter	7.46
CLIBASIA_03625	kup	Putative potassium uptake transport system protein	1.42
CLIBASIA_04115	NA	Kef-type K^+ transport system, predicted NAD-binding component	3.62
CLIBASIA_04145a	nodT	Nodulation outer membrane efflux protein	1.54
CLIBASIA_04415	NA	Putative threonine efflux protein	2.09
CLIBASIA_04810	NA	ABC transporter nucleotide binding/ATPase	−1.26
CLIBASIA_05070	NA	ABC-type amino acid transport/signal transduction systems, periplasmic component/domain	2.97
CLIBASIA_05075	NA	ABC-type amino acid transport system, permease component	7.96
CLIBASIA_05125	NA	ATPase components of ABC transporters with duplicated ATPase domains	4.80

• ^a Psyllid RNA samples were extracted from a single psyllid (see Experimental procedures).
• ^b NA, gene name was not assigned.
• ^c Information on gene product was based on the annotation of the genome of Ca. L. asiaticus strain psy62 and definition in the KEGG database.
• ^d Fold change (log₂ ratio) is the relative gene expression (in planta versus in psyllid) of Ca. L. asiaticus. Positive value indicates that gene was overexpressed in planta compared with in psyllid; negative value indicates that gene was overexpressed in psyllid compared with in planta.

Table 3. Expression profile of genes involved in metabolism pathway of Candidatus Liberibacter asiaticus in planta compared with in psyllid

Locus tag	Gene nameb	Gene productc	KEGG pathway	Relative expressiond
Carbohydrate metabolism
CLIBASIA_00375	fumC	Fumarate hydratase	Citrate cycle	3.39
CLIBASIA_00825	glk	Glucokinase	Glycolysis	2.78
CLIBASIA_01065	NA	UDP-glucose 4-epimerase	Galactose	3.44
CLIBASIA_01165	NA	5-Amino-6-(5-phosphoribosylamino)uracil reductase/diaminohydroxyphosphoribosylaminopyrimidine	Riboflavin metabolism	3.88
CLIBASIA_01680	acnA	Aconitate hydratase	Glyoxylate, dicarboxylate metabolism/citrate cycle	5.27
CLIBASIA_02695	NA	Fructose-bisphosphate aldolase	Glycolysis	3.4
CLIBASIA_02700	pgk	Phosphoglycerate kinase	Glycolysis	2.1
CLIBASIA_02705	NA	Glyceraldehyde 3-phosphate dehydrogenase	Glycolysis	2.77
CLIBASIA_02710	tkt	Transketolase	Pentose phosphate pathway	6.24
CLIBASIA_02785	eno	Phosphopyruvate hydratase	Glycolysis	2.46
CLIBASIA_02810	lpdA	Dihydrolipoamide dehydrogenase	Glycolysis	1.49
CLIBASIA_03595a	pckA	Phosphoenolpyruvate carboxykinase	Glycolysis	0.79
CLIBASIA_04750	mdh	Malate dehydrogenase	Glyoxylate,dicarboxylate metabolism	2.32
CLIBASIA_05045	NA	Phosphoglucomutase	Glycolysis	3.47
Energy metabolism
CLIBASIA_00340	glnA	Glutamine synthetase protein	Alanine, aspartate and glutamate metabolism	4.76
CLIBASIA_00345	glsA	Glutaminase	d-glutamine and d-glutamate metabolism	1.76
CLIBASIA_00585a	NA	Inorganic pyrophosphatase	Oxidative phosphorylation	2.2
CLIBASIA_02750	NA	Carbonate dehydratase	Nitrogen metabolism	1.33
CLIBASIA_03735	nuoA	NADH dehydrogenase subunit A	Oxidative phosphorylation	−2.09
CLIBASIA_04725	sdhA	Succinate dehydrogenase flavoprotein subunit	Oxidative phosphorylation	2.47
Nucleotide metabolism
CLIBASIA_00055	gmk	Guanylate kinase	Purine metabolism	2.41
CLIBASIA_00400	pyrG	CTP synthetase		1.51
CLIBASIA_00515	tmk	Thymidylate kinase	Pyrimidine metabolism	3.87
Metabolism of terpenoids and polyketides
CLIBASIA_03210	ispB	Octaprenyl-diphosphate synthase protein	Terpenoid backbone biosynthesis	2.23
CLIBASIA_04105	rfbA	Glucose-1-phosphate thymidylyltransferase	Polyketide sugar unit biosynthesis	1.82
CLIBASIA_05065	NA	Geranyltranstransferase protein	Terpenoid backbone biosynthesis	3.65
Metabolism of cofactors and vitamins
CLIBASIA_00425	hemH	Ferrochelatase	Porphyrin and chlorophyll metabolism	3.29
CLIBASIA_01160	NA	Riboflavin synthase subunit α	Riboflavin metabolism	−2.99
CLIBASIA_01965	folD	Methylenetetrahydrofolate dehydrogenase/cyclohydrolase protein	One-carbon pool by folate	1.9
CLIBASIA_02815	NA	Lipoyl synthase	Lipoic acid metabolism	2.84
CLIBASIA_03515	hemE	Uroporphyrinogen decarboxylase	Porphyrin and chlorophyll metabolism	3.93
CLIBASIA_03560	folC	FolC bifunctional protein	Folate biosynthesis	2.49
CLIBASIA_04185	acpS	4′-Phosphopantetheinyl transferase	Pantothenate and CoA biosynthesis	1.84
CLIBASIA_04670	NA	Hypothetical protein	Folate biosynthesis	2.65
CLIBASIA_04685	hemD	Uroporphyrinogen-III synthase	Porphyrin and chlorophyll metabolism	3.03
CLIBASIA_04695	hemC	Porphobilinogen deaminase	Porphyrin and chlorophyll metabolism	4.1
CLIBASIA_04875	NA	Coproporphyrinogen III oxidase	Porphyrin and chlorophyll metabolism	2.16
Amino acid metabolism
CLIBASIA_04835	kamA	l-lysine 2,3-aminomutase protein	Lysine degradation	6.39
CLIBASIA_02600	NA	Creatinine amidohydrolase	Arginine/proline metabolism	8.43
CLIBASIA_00220	NA	Phytoene synthase protein	Biosynthesis of secondary metabolites	8.56
CLIBASIA_03995	thrA	Homoserine dehydrogenase	Glycine, serine and threonine metabolism	2.30
CLIBASIA_01170	glyA	Serine hydroxymethyltransferase	Glycine, serine and threonine metabolism	2.32
CLIBASIA_04210	glmS	Glucosamine-fructose-6-phosphate aminotransferase	Alanine, aspartate and glutamate metabolism	−1.71
Metabolism of other amino acids
CLIBASIA_02780	kdsA	2-Dehydro-3-deoxyphosphooctonate aldolase	Lipopolysaccharide biosynthesis	1.82
CLIBASIA_02915	gshB	Glutathione synthetase	Glutathione metabolism	1.59
CLIBASIA_03015	gnd	6-Phosphogluconate dehydrogenase	Glutathione metabolism	7.01
CLIBASIA_03280	kdsB	3-Deoxy-manno-octulosonate cytidylyltransferase	Lipopolysaccharide biosynthesis	3.29
CLIBASIA_03290	lpxB	Lipid-A-disaccharide synthase	Lipopolysaccharide biosynthesis	6.12
CLIBASIA_03420	mviN	Integral membrane protein MviN	Peptidoglycan biosynthesis	3.62
CLIBASIA_04785	gor	Glutathione reductase	Glutathione metabolism	4.17
Genetic information processing
Folding, sorting and degradation
CLIBASIA_03720	NA	Chaperonin GroEL	RNA degradation	1.85
Replication and repair
CLIBASIA_02530	NA	Primosome assembly protein PriA	Homologous recombination	4.97

• ^a Psyllid RNA samples were extracted from a single psyllid (see Experimental procedures).
• ^b NA, gene name was not assigned.
• ^c Information on gene product was based on the annotation of the genome of Ca. L. asiaticus strain psy62 and definition in the KEGG database.
• ^d Fold change (log₂ ratio) is the relative gene expression (in planta versus in psyllid) of Ca. L. asiaticus. Positive value indicates that gene was overexpressed in planta compared with in psyllid; negative value indicates that gene was overexpressed in psyllid compared with in planta.

Transport systems

A large number of genes involved in active transport were overexpressed in planta (Table 2). ABC transporters are known to be involved in the virulence of various bacteria, and the virulence is associated with the uptake of nutrients (Darwin and Miller, 1999), uptake of metal ions (Boyer et al., 2002) or cell attachment (Tamura et al., 2002). Candidatus L. asiaticus encodes a much larger number of ABC transporters compared with other intracellular bacteria, which suggests that they may be involved in virulence or the elicitation of symptoms (Duan et al., 2009). In our study, the expression of 16 ABC transporter genes was up-regulated in planta. In psyllid, however, only two ABC transporter genes were up-regulated compared with in planta (Table 2).

Many of the ABC transporter genes up-regulated in planta were involved in the uptake or efflux of essential micronutrients, in order to maintain appropriate levels of these nutrients in the cell (Table 2). The genes CLIBASIA_02120 and CLIBASIA_02125, encoding homologous components of the SitABCD system responsible for high-affinity uptake of Mn²⁺ and Fe²⁺ (Zhou et al., 1999), were up-regulated in planta. The SitABCD transport system has been shown to be essential for the virulence of various pathogenic bacteria, including Salmonella enterica serovar Typhimurium and Escherichia coli (Boyer et al., 2002; Sabri et al., 2008). Iron is essential to almost all organisms, including bacteria. It participates in many major biological processes, including respiration, the trichloroacetic acid cycle, oxygen transport, gene regulation and DNA biosynthesis (Andrews et al., 2006). Iron uptake is important in the plant–bacterium interaction in the symbiont S. meliloti and the pathogen A. tumefaciens (Chao et al., 2005; Kitphati et al., 2007). The ability to acquire Mn²⁺ plays a major role in virulence, and also contributes to protection against oxidative stress in various plant and animal pathogens (Li et al., 2011; Papp-Wallace and Maguire, 2006). Studies have also shown that sitABCD genes are induced specifically during the systemic infection of Salmonella enterica serovar Typhimurium. During infection, the host has been reported to actively reduce the availability of extracellular iron, as a nonspecific defence mechanism, prompting successful pathogens to evolve efficient strategies to acquire iron from this iron-limiting environment (Wooldridge and Williams, 1993), including the utilization of low-iron-induced genes, such as sitABCD (Janakiraman and Slauch, 2000).

Our results also showed the up-regulation of the pstSCAB-phoU operon (CLIBASIA_02950, CLIBASIA_02955, CLIBASIA_02960, CLIBASIA_02965, CLIBASIA_02970) encoding an ABC-type transporter system for phosphate uptake into the bacterial cell of Ca. L. asiaticus in the host plant (Tables 1 and 2). Phosphate plays a major role in the conversion and transfer of energy in the tricarboxylic acid cycle and in glycolysis. Several studies have associated the pst-phoU system with the survival and virulence of bacteria, with mutations causing reduced virulence, sensitivity to the bactericidal effect of serum, reduction in the amount of capsular antigen at the cell surface, impaired colonization ability and attachment, and reduced capacity to multiply within phagocytes and serum, collectively suggesting that bacterial cell surface modifications occur in the mutants, and implicating the pst-phoU system in the regulation of bacterial pathogenicity (Lamarche et al., 2008). In addition to the role in phosphate uptake, PhoU has also been found to be a master regulator in a process of persistence involved in bacterial survival to antibiotic treatment in E. coli and Mycobacterium tuberculosis. Mutation of phoU leads to higher susceptibility to diverse stresses, including antibiotics, starvation, weak acid, heat and energy inhibitors (Li and Zhang, 2007; Shi and Zhang, 2010).

The gene proX (CLIBASIA_01135), involved in glycine betaine/proline transport, was up-regulated in planta compared with in psyllid (Table 2). The induction of proX in planta might contribute to the adaptation of Ca. L. asiaticus in the two diverse environments of the plant and insect systems. After inoculation into the plant phloem by the insect, the pathogen encounters a change in osmolarity and must protect itself from dehydration and loss of turgor. Many prokaryotes deal with a change in osmolarity through the uptake of compatible solutes, which do not interfere with the metabolism of the organism even at high concentrations (Roebler and Muller, 2001). One of the most common osmoprotectants is glycine betaine, utilized by many bacteria, including E. coli (Perroud and Le Rudulier, 1985). In addition to being an osmoprotectant, glycine betaine and proline betaine can also function as carbon, nitrogen and energy sources supporting growth in certain bacteria, such as Rhizobium sp. (Bernard et al., 1986).

Metabolic pathway

In our study, eight genes (CLIBASIA_00825, CLIBASIA_02695, CLIBASIA_02700, CLIBASIA_02705, CLIBASIA_02785, CLIBASIA_02810, CLIBASIA_03595 and CLIBASIA_05045), encoding enzymes involved in glycolysis, were up-regulated in the host plant (Table 3). Candidatus L. asiaticus is known to survive exclusively in the citrus phloem where glucose and other nutrients, derived from photosynthesis, are transported. Furthermore, elevated glucose was observed in Ca. L. asiaticus-infected citrus (Fan et al., 2010). The high expression of glycolysis-associated genes in planta indicates that Ca. L. asiaticus can use glucose acquired from the host plant for the generation of energy to support the intracellular growth of the pathogen in the plant. In addition, the glycolysis pathway is also up-regulated during infection and is important to the virulence of Yersinia pseudotuberculosis and Salmonella enterica serovar Typhimurium (Chaudhuri et al., 2009; Rosso et al., 2008).

kdsA (CLIBASIA_02780), kdsB (CLIBASIA_03280) and lpxB (CLIBASIA_03290), involved in lipopolysaccharide (LPS) biosynthesis, together with one LPS ABC transporter gene, lptB (CLIBASIA_03155), were overexpressed in planta (Tables 2 and 3). The up-regulation of LPS genes might be important for Ca. L. asiaticus to survive in planta, as LPS is the major component of the outer membrane in Gram-negative bacteria which protects bacterial cells from unfavourable plant environments (Dow et al., 1995). In plant–pathogen interactions, LPS also functions as a pathogen-associated molecular pattern (PAMP) by eliciting basal defence-related responses (Parker, 2003). The infection of Ca. L. asiaticus was observed to induce the expression of many pathogenesis-related genes, including WRKY4, WRKY6, ERF-1 and ERF-2, in host citrus (sweet orange, C. sinensis), indicating an activation of defence mechanisms (Kim et al., 2009). Phytopathogenic bacteria have evolved the ability to deliver effector molecules inside the host cell via T3SS to suppress plant defences (Jones and Dangl, 2006). However, no T3SS was found in the genome of Ca. L. asiaticus. How Ca. L. asiaticus suppresses plant defence remains to be determined.

In this study, genes involved in haem biosynthesis, including hemH (CLIBASIA_00425), hemE (CLIBASIA_03515), hemC (CLIBASIA_04695) and hemD (CLIBASIA_04685), were up-regulated in planta (Table 3). Previous studies have shown that physiological factors, such as oxygen, nitrate and carbon sources, act as signals for the regulation of haem biosynthesis (Schobert and Jahn, 2002). Haem, a biological catalyst synthesized by bacteria, functions not only as a source of iron for bacterial growth, but also as a regulator involved in virulence regulation (Wandersman and Delepelaire, 2004). Haem is also involved in various aspects of oxidative metabolism, including oxidative stress responses, oxygenation reactions and detoxification (Panek and O'Brian, 2002). hemH encodes a ferrochelatase involved in the last step of the haem metabolic pathway and catalyses the insertion of a ferrous iron atom into the porphyrin ring. Ferrochelatase has been shown previously to be essential for intracellular survival and virulence (Almiron et al., 2001). The genes hemE, hemD and hemC, encoding uroporphyrinogen decarboxylase, uroporphyrinogen-III synthase and porphobilinogen deaminase, respectively, are also involved in bacterial haem biosynthesis (Frankenberg et al., 2003). In P. aeruginosa, these three genes have been shown to be involved indirectly in virulence by regulating the secretion of a known virulence factor, the exopolysaccharide, alginate (Mohr et al., 1994). Similarly, the induction of haem biosynthesis genes in planta may be important in the survival and virulence of Ca. L. asiaticus.

Secretion system

The majority of bacterial secretory proteins and membrane proteins are translocated in a Sec-dependent secretion system. Genome sequencing has revealed that Ca. L. asiaticus harbours all of the basic components of the Sec machinery (Duan et al., 2009). In this work, the expression of secE (CLIBASIA_00140) and secD/F (CLIBASIA_04120) was found to be up-regulated in planta compared with in psyllid (Table 4). SecE is the component of the SecYEG translocase complex which facilitates the translocation and membrane insertion of the majority of inner membrane proteins (Dalbey and Chen, 2004). SecD/F is one of the translocase subunits, functionally associated with the SecYEG complex by optimizing the secretion of substrate proteins (Duong and Wickner, 1997). The up-regulation of Sec-associated genes in planta is interesting because the Sec-dependent secretory pathway is involved in the virulence of various intracellular pathogenic bacteria via the secretion of various proteins (including virulence factors) that play important roles in the host–pathogen interaction (Van Gijsegem et al., 1993). The absence of other well-known effector secretion systems, including T3SS, makes the Sec-dependent secretion pathway more important for the virulence of Ca. L. asiaticus, as it may play a major role in the translocation of virulence factors.

Table 4. Expression profile of genes encoding secretion system of Candidatus Liberibacter asiaticus in planta compared with in psyllid

Locus tag	Gene nameb	Gene productc	Relative expressiond
CLIBASIA_00070a	NA	Protein with a putative signal peptide	−1.37
CLIBASIA_00140	secE	Preprotein translocase subunit involved in bacterial secretion system	4.84
CLIBASIA_00215	NA	Protein with a putative signal peptide	1.15
CLIBASIA_00470	NA	Protein with a putative signal peptide	7.36
CLIBASIA_00525a	NA	Protein with a putative signal peptide	−2.40
CLIBASIA_00530a	NA	Protein with a putative signal peptide	−0.84
CLIBASIA_01345	NA	Serralysin. RTX toxins and related Ca^2+-binding proteins	5.55
CLIBASIA_02145a	NA	Protein with a putative signal peptide	1.86
CLIBASIA_02215a	NA	Protein with a putative signal peptide	2.59
CLIBASIA_02275	NA	Protein with a putative signal peptide	1.77
CLIBASIA_02610a	NA	Protein with a putative signal peptide	−2.75
CLIBASIA_02845	NA	Protein with a putative signal peptide	4.17
CLIBASIA_02990	ffh	Signal recognition particle protein involved in bacterial secretion system	1.81
CLIBASIA_03085	NA	Protein with a putative signal peptide	4.94
CLIBASIA_03180	lspA	Lipoprotein signal peptidase involved in bacterial secretion system	1.46
CLIBASIA_03230	NA	Protein with a putative signal peptide	3.52
CLIBASIA_03915a	NA	Protein with a putative signal peptide	2.75
CLIBASIA_04025a	NA	Protein with a putative signal peptide	3.40
CLIBASIA_04030a	NA	Protein with a putative signal peptide	0.83
CLIBASIA_04040	NA	Protein with a putative signal peptide	2.43
CLIBASIA_04120	secD/F	Bifunctional preprotein translocase subunit involved in bacterial secretion system	4.33
CLIBASIA_04190	lepB	Type I signal peptidase involved in bacterial secretion system	1.46
CLIBASIA_04330a	NA	Protein with a putative signal peptide	1.00
CLIBASIA_04515	NA	Protein with a putative signal peptide	4.03
CLIBASIA_04530a	NA	Protein with a putative signal peptide	2.67
CLIBASIA_04540a	NA	Protein with a putative signal peptide	−0.91
CLIBASIA_04560a	NA	Protein with a putative signal peptide	2.04
CLIBASIA_04580	NA	Protein with a putative signal peptide	2.04
CLIBASIA_05150a	NA	Protein with a putative signal peptide	1.70
CLIBASIA_05320a	NA	Protein with a putative signal peptide	4.39
CLIBASIA_05480	NA	Protein with a putative signal peptide	7.67
CLIBASIA_05570	NA	Protein with a putative signal peptide	6.79

• ^a Psyllid RNA samples were extracted from a single psyllid (see Experimental procedures).
• ^b NA, gene name was not assigned.
• ^c Information on gene product was based on the annotation of the genome of Ca. L. asiaticus strain psy62 and definition in the KEGG database. Signal peptide was predicted using signal V4.0 program.
• ^d Fold change (log₂ ratio) is the relative gene expression (in planta versus in psyllid) of Ca. L. asiaticus. Positive value indicates that gene was overexpressed in planta compared with in psyllid; negative value indicates that gene was overexpressed in psyllid compared with in planta.

Proteins secreted by the Sec-dependent pathway are characterized by the presence of an N-terminal signal peptide, indicating potential translocation into the host. In this study, 26 hypothetical genes with a predicted N-terminal signal peptide showed altered expression in the host plant or insect. Among the 26 genes, 21 were up-regulated in planta (Table 4). Specifically, the expression of CLIBASIA_05480, CLIBASIA_00470 and CLIBASIA_05570 was increased dramatically by a factor of more than six (log₂ value, equal to 64 times fold change) in planta compared with in psyllid. Although no known functional domain was found, the significant induction of these secreted proteins indicates that they may be important in the plant–pathogen interaction between Ca. L. asiaticus and the host plant.

The gene product of CLIBASIA_05150 harbours an ‘OMP_b-brl’ domain (amino acids 52–225, e value = 6.70E-5) and a transmembrane segment predicted by the TMHMM2 program. The OMP_b-brl domain assumes a membrane-bound β-barrel which is conserved in a wide range of outer membrane proteins, such as OmpA, OmpX and NspA. Proteins with this domain have been suggested to be involved in diverse biological functions, including virulence, membrane stability and resistance to environmental stresses (Kim et al., 2010; Ried and Henning, 1987; Wang, 2002). In particular, many of the outer membrane proteins with a barrel structure are involved in the initial interaction between pathogenic bacteria and their hosts (McClean, 2012). For example, OmpA was found to be the major protein of Cronobacter sakazakii, binding to fibronectin of human cells, which is the first step in the invasion of the host by the pathogen (Mohan Nair et al., 2009).

Five of the 26 putative secreted proteins were up-regulated in the psyllid host. One, the gene CLIBASIA_02610, encoding a hypothetical protein, harbours an imelysin domain (amino acids 37–391, e value = 3.2E-75) (Table 4). Two biological functions have been assigned to imelysin: (i) peptide cleavage as a metalloproteinase; (ii) iron uptake. Imelysin was first named in P. aeruginosa as a zinc peptidase involved in insulin cleavage (Fricke et al., 1999). A homologous peptidase was found in the nematode pathogen Xenorhabdus nematophila, in which the imelysin homologue was indicated to be involved in insect immunosuppression by destroying antibacterial factors present in insect haemolymph (Caldas et al., 2002). In addition, imelysin is important in the support of bacterial growth of Synechococcus sp. in iron-limited conditions and is involved in iron uptake or metabolism in P. aeruginosa and V. cholera (Reddy et al., 1988; Xu et al., 2011). Thus, the up-regulation of the gene CLIBASIA_02610 in psyllid may indicate that this gene is important in the survival and propagation of Ca. L. asiaticus in the insect host.

In addition to the Sec-dependent pathway, Ca. L. asiaticus also contains an intact type I secretion system (T1SS) (Duan et al., 2009). It is known that offensive enzymes and effectors can be secreted via T1SS in plant and animal pathogenic bacteria (Van Gijsegem et al., 1993). A putative T1SS effector, a serralysin, encoded by CLIBASIA_01345, which is located next to the T1SS locus in the genome, has been identified recently by computational analysis of Ca. L. asiaticus (Cong et al., 2012). In this work, we found that the expression of CLIBASIA_01345 was up-regulated in planta compared with in psyllid (Table 4). Serralysin is a secreted metalloprotease produced by a wide range of microorganisms, including plant and human pathogenic bacteria, such as Serratia marcescens, P. aeruginosa, Erwinia chrysanthemi, Proteus mirabilis and Caulobacter crecentus (Dahler et al., 1990; Maeda and Morihara, 1995). It has been shown that serralysin inactivates diverse antimicrobial proteins and peptides (Schmidtchen et al., 2002). For example, serralysin produced by P. mirabilis has been reported to degrade host immunoglobulins and cleave antimicrobial peptides, including human β-defensin and LL-37 (Belas et al., 2004). These allow P. mirabilis to modify the host immune response. The production of antimicrobial proteins and peptides is one of the major defence strategies utilized by plants in response to infection by pathogenic organisms (Castro and Fontes, 2005). The up-regulation of the serralysin biosynthesis gene in planta indicates that Ca. L. asiaticus may also utilize serralysin to modify plant defence, possibly by degrading host antimicrobial peptides. It has also been suggested that serralysin might aid in the acquisition of carbon and nitrogen for bacterial growth and metabolism through the proteolysis of host proteins and nutrient uptake (Basu and Apte, 2008; Belas et al., 2004). Serralysin may further help Ca. L. asiaticus to survive in its hosts. In addition, the introduction of exogenous antimicrobial peptides into citrus plants, by various transgenic approaches, is being used to control HLB. The presence of serralysin poses a potential challenge for the selection of efficient antimicrobial peptides against Ca. L. asiaticus. Thus, the serralysin of Ca. L. asiaticus could be a potential target for the screening of antimicrobial compounds for the control of HLB.

Flagellar assembly

The expression of genes involved in flagellar assembly, including fliF (CLIBASIA_02910), flgI (CLIBASIA_01305) and flgD (CLIBASIA_02035), and the motB (CLIBASIA_02080) gene involved in motor function, was up-regulated in planta. In contrast, flgL (CLIBASIA_02050), flgK (CLIBASIA_02055) and fliE (CLIBASIA_01320) were overexpressed in psyllid (Table 5). In spite of the small genome size, Ca. L. asiaticus has retained most of its flagellar genes (Duan et al., 2009), although electron microscopy studies have failed to detect the presence of flagella associated with the bacterium in the phloem (Bové, 2006). Studies have shown that, in intracellular pathogens, the presence of flagella confers a growth disadvantage (Macnab, 1996) and is energetically expensive for the bacteria, unless it serves another essential purpose in the plant or insect, such as the export of proteins including virulence factors (Young et al., 1999). The absence of flagellar genes is common in many nonmotile obligate intracellular endosymbionts, such as Blochmannia floridanus and Baumannia (Gil et al., 2003; Wu et al., 2006). Furthermore, Ca. L. asiaticus lacks a functional T3SS, and hence it is possible that the flagellar system may be involved in the delivery of virulence determinants inside the host cell (Duan et al., 2009). This was observed in Buchnera aphidicola, a nonmotile endosymbiont which has lost most of the flagellar assembly genes, whereas numerous basal bodies (lacking the filament part of the flagellum) covered the cell surface in this pathogen, possibly involved in protein export to the host (Maezawa et al., 2006). Interestingly, the expression of the Ca. L. asiaticus flagellin gene, CLIBASIA_02090, was significantly lower in planta compared with in psyllid (Table 5), which might be a tactic used by the pathogen to avoid the elicitation of plant defence responses. Flagellin has been shown to trigger the host response and to induce basal defence in planta as a PAMP factor (Felix et al., 1999).

Table 5. Expression profile of motility-related genes of Candidatus Liberibacter asiaticus in planta compared with in psyllid

Locus tag	Gene nameb	Gene productc	Relative expressiond
CLIBASIA_02910	fliF	Flagellar MS-ring protein involved in flagellar assembly	3.39
CLIBASIA_02035	flgD	Flagellar basal body rod modification protein involved in flagellar assembly	1.96
CLIBASIA_02055	flgK	Flagellar hook-associated protein FlgK	−2.03
CLIBASIA_02080	motB	Flagellar motor protein MotB involved in flagellar assembly	4.86
CLIBASIA_02090	NA	Flagellin domain-containing protein	−2.59
CLIBASIA_02050a	flgL	Flagellar hook-associated protein FlgL	−1.28
CLIBASIA_03105	NA	Flp/Fap pilin component	−4.05
CLIBASIA_03945a	cheY	Probable two-component response regulator protein involved in bacterial chemotaxis	2.99
CLIBASIA_01305	flgI	Flagellar basal body P-ring protein involved in flagellar assembly	3.42
CLIBASIA_01320a	fliE	Flagellar hook-basal body protein FliE	−2.29

• ^a Psyllid RNA samples were extracted from a single psyllid (see Experimental procedures).
• ^b NA, gene name was not assigned.
• ^c Information on gene product was based on the annotation of the genome of Ca. L. asiaticus strain psy62 and definition in the KEGG database.
• ^d Fold change (log₂ ratio) is the relative gene expression (in planta versus in psyllid) of Ca. L. asiaticus. Positive value indicates that gene was overexpressed in planta compared with in psyllid; negative value indicates that gene was overexpressed in psyllid compared with in planta.

Environmental stress

In addition to providing the nutrients for the pathogen, plants also produce toxic chemicals and pathogen-degrading enzymes, and undergo deliberate cell suicide in response to invasion by a pathogen (Freeman and Beattie, 2008). Interestingly, the expression of ATP-dependent protease genes, including clpX (CLIBASIA_00780) and ftsH (CLIBASIA_02160), which are involved in the adaptation of the bacterium to environmental stresses, and are implicated in the virulence of some pathogens (Lithgow et al., 2004), was up-regulated in planta (Table 6). In addition, known stress response-related genes dnaJ (CLIBASIA_02625), grpE (CLIBASIA_02945) and groEL (CLIBASIA_03720) (Farr and Kogoma, 1991; Gomes and Simão, 2009) were also overexpressed in planta (Tables 1 and 6). In E. coli, the genes clpX, ftsH, grpE, groEL and dnaJ belong to the heat-shock regulon controlled by alternative sigma factor RpoH (Nonaka et al., 2006), which is required for stress resistance and environment fitness. The up-regulation of clpX, ftsH, grpE, groEL and dnaJ was consistent with the induction of rpoH in the host plant (Table 1). In addition, the expression of gshB (CLIBASIA_02915), gor (CLIBASIA_04785) and gnd (CLIBASIA_03015), encoding enzymes involved in the metabolism of glutathione, was also up-regulated in the host plant (Table 3). Glutathione peroxidase is a common antioxidant of bacteria for efficient protection against oxidative damage (Arenas et al., 2011). The up-regulation of these stress resistance-related genes is believed to protect Ca. L. asiaticus from harmful plant environments and to contribute to successful colonization of the pathogen in the host plant.

Table 6. Expression profile of genes not classified in the KEGG pathway or other functional category of Candidatus Liberibacter asiaticus in planta compared with in psyllid

Locus tag	Genea	Gene productb	Relative expressionc
CLIBASIA_00495	NA	Metal-dependent hydrolases of the β-lactamase superfamily I	6.12
CLIBASIA_00780	clpX	ATP-dependent protease ATP-binding subunit	3.05
CLIBASIA_00970	NA	ComF family protein. Predicted amidophosphoribosyltransferases	2.98
CLIBASIA_01240	fpr	Flavodoxin reductases (ferredoxin-NADPH reductases) family	2.02
CLIBASIA_01265	etfB	Electron transfer flavoprotein, β subunit	2.13
CLIBASIA_01480	NA	Prolipoprotein diacylglyceryltransferase	4.17
CLIBASIA_02160	ftsH	Metalloprotease. ATP-dependent Zn proteases	3.93
CLIBASIA_02605	NA	tRNA and rRNA cytosine-C5-methylases. NOL1/NOP2/SUN family signature protein	1.59
CLIBASIA_02625	dnaJ	DnaJ-class molecular chaperone with C-terminal Zn finger domain	2.80
CLIBASIA_02850	NA	Membrane proteins related to metalloendopeptidases	2.04
CLIBASIA_02945	grpE	Molecular chaperone GrpE (heat shock protein)	1.45
CLIBASIA_03135	NA	Uncharacterized protein, similar to the N-terminal domain of Lon protease	7.59
CLIBASIA_03170	NA	Periplasmic serine proteases (ClpP class)	2.62
CLIBASIA_03175	NA	Bacterial nucleoid DNA-binding protein. Integration host factor, β subunit	1.60
CLIBASIA_03315	omp	Surface antigen (D15).Outer membrane protein/protective antigen OMA87	2.24
CLIBASIA_03490	gidB	Glucose-inhibited division protein B	1.82
CLIBASIA_04070	NA	Oligoendopeptidase F	2.93
CLIBASIA_04290	NA	Putative hydrolase serine protease transmembrane protein	3.80
CLIBASIA_04830	NA	FeS assembly scaffold SufA	1.74
CLIBASIA_04885	NA	Alanyl-tRNA synthetase	4.12
CLIBASIA_05000	NA	Cell division protein FtsW	5.46
CLIBASIA_05405	NA	Outer membrane assembly lipoprotein YfiO	3.69

• ^a NA, gene name was not assigned.
• ^b Information on gene product was based on the annotation of the genome of Ca. L. asiaticus strain psy62 and definition in the KEGG database.
• ^c Fold change (log₂ ratio) is the relative gene expression (in planta versus in psyllid) of Ca. L. asiaticus. Positive value indicates that gene was overexpressed in planta compared with in psyllid; negative value indicates that gene was overexpressed in psyllid compared with in planta.

Hypothetical proteins

Large numbers of hypothetical genes of Ca. L. asiaticus were found in this work whose expression was changed significantly in planta or in psyllid (Table S3, see Supporting Information). Information on the biological function played by these hypothetical genes is limited. Domain analysis revealed that 10 genes encoded a protein with a transmembrane helix domain, indicating a putative role in membrane-associated function. Specifically, the genes CLIBASIA_01365, CLIBASIA_04165 and CLIBASIA_05050, harbouring a von Willebrand factor type A (vWFA) domain, were up-regulated by a factor of more than three (log₂ value, equal to eight times fold change) in planta compared with in psyllid. vWFA is widely distributed in eukaryotes and prokaryotes. It functions as an adhesive glycoprotein on the surface of blood cells in mammals and is involved in multiple processes, including adhesion, migration and signal transduction, by interacting with a large array of ligands (Colombatti et al., 1993). In contrast, the functions of the majority of bacterial vWFA are still unknown, with only a few exceptions. For example, the TerY protein of E. coli has been found to protect bacterial cells from the toxic effects of heavy metals (Whelan et al., 1997), implying a binding role to metal ions (Ponting et al., 1999). The vWFA domain-containing D subunit of the enzyme magnesium-protoporphyrin IX chelatase interacts with the enzyme's I subunit in Synechocystis PCC6803 (Jensen et al., 1998). Similarly, the vWFA proteins of Ca. L. asiaticus might contribute to the compatible interactions between the pathogen and host plant by interacting with putative ligands of the plant.

Why are so many genes up-regulated in planta compared with in psyllid?

The gene expression profiles revealed that 198 genes of Ca. L. asiaticus showed a significant change in expression in planta compared with in psyllid. It is interesting that most (182 of 198 genes, 92%) were up-regulated in planta compared with in psyllid. One of the reasons could be the putative bias in the choice of the target genes in our study. As described above, the 381 genes tested in this work were mainly associated with virulence and/or survival, which only represents 32% of the total predicted genes in the whole genome of Ca. L. asiaticus. It is possible that some of the genes activated in psyllid were not included in this assay. Another possible reason is that plants are more favourable than psyllids for the gene expression of Ca. L. asiaticus. This probably results from the co-evolution among Ca. L. asiaticus, citrus and psyllids. It is probable that Ca. Liberibacter spp. evolved from an ancestor in the Rhizobiaceae family through adaptive, diversifying and reductive evolutionary processes that occurred during host adaptation (Toft and Andersson, 2010). This is possibly a result of the intimate relationship between rhizobia and plant roots (Gage, 2004). The intimate associations of Ca. Liberibacter spp. with plants as endophytes predispose them to frequent encounters with herbivorous insects, providing Ca. Liberibacter spp. with ample opportunity to colonize and eventually evolve alternative associations with insects (Nadarasah and Stavrinides, 2011). The genetic contents and regulation of Ca. L. asiaticus are thus more suitable for its interaction with plants than with insects, e.g. suppression of plant defence and acquisition of nutrients. As a successful pathogen, Ca. L. asiaticus has gained the ability to overcome the plant defence mechanism. It is consistent with our results that the genes of Ca. L. asiaticus up-regulated in planta are mainly involved in the virulence and/or survival of the pathogen.

Conclusions

In this study, transcriptional profiling of the genes involved in the survival and virulence of Ca. L. asiaticus was investigated in two of its hosts, sweet orange ‘Valencia’ and the psyllid vector D. citri. Our study indicated that Ca. L. asiaticus alters its expression of the genes involved in transcriptional regulation, transport systems, metabolic pathways, secretion systems and stress resistance in a host-specific manner to adapt to the distinct environment of plant and insect. To our knowledge, this is the first large-scale study to evaluate the differential expression of Ca. L. asiaticus genes in a plant host and its insect vector.

Experimental Procedures

Plant materials

Nine-month-old seedlings of sweet orange (C. sinensis) ‘Valencia’ infected with Ca. L. asiaticus were used to evaluate the differential gene expression of Ca. L. asiaticus in infected plants. The plants were graft inoculated with two pieces of Ca. L. asiaticus-infected budwood from PCR-positive HLB source trees. Inoculated plants were kept in a US Department of Agriculture-Animal and Plant Health Inspection Service (USDA-APHIS)-approved secure glasshouse, with a temperature ranging from 26 to 32 °C at the Citrus Research and Education Center, University of Florida, Lake Alfred, FL, USA. After 3 months of graft inoculation, infection was verified by PCR assays using the specific primers A2-J5 and CQULA04F-CQULA04R as described previously (Hocquellet et al., 1999; Wang et al., 2006) (data not shown).

Psyllid materials

Healthy psyllids (D. citri) were obtained from a healthy culture of psyllids which was established from field populations in Polk Co., FL, USA (28.0′N, 81.9′W) during 2000, prior to the discovery of HLB in Florida. Psyllids were maintained at 27 ± 1 °C, 80% ± 2% relative humidity and a 16 h : 8 h light : dark photoperiod without exposure to insecticides. Candidatus L. asiaticus-infected psyllids were maintained on confirmed Ca. L. asiaticus-infected sweet orange in secure, insect-proof enclosures at the Citrus Research and Education Center, Lake Alfred, FL, USA.

Extraction of total RNA from infected plant and psyllid

Total RNA was extracted using the RNeasy Plant Mini Kit (Qiagen, Valencia, CA, USA), according to the manufacturer's instructions. For the extraction of RNA from Ca. L. asiaticus-infected plants, the midrib of symptomatic leaves was collected and ground in liquid nitrogen to a fine powder before RNA isolation. Two methods were used for RNA isolation from putative Ca. L. asiaticus-infected psyllids. As a validation of the gene expression profile, the psyllid RNA samples isolated by these two methods were used for QRT-PCR analysis. The first method was performed by pooling psyllids before RNA extraction. As a result of the lack of visible differences between Ca. L. asiaticus-infected and uninfected psyllids, and the low titre of Ca. L. asiaticus in psyllids (Benyon et al., 2008), 15–20 psyllids were pooled and ground in liquid nitrogen before RNA isolation. The other method involved the extraction of RNA from a single psyllid. The single psyllid was homogenized in 600 μL of lysis buffer using a TissueRuptor (Qiagen). Only the clear supernatant from the lysis was used for RNA extraction following the manufacturer's instructions. In this work, we extracted RNA from a total of 210 psyllids individually. These 210 psyllid RNA samples were subjected to QRT-PCR assay to test for infection of Ca. L. asiaticus using primers CQULA04F-CQULA04R, which specifically target the β-operon of Ca. L. asiaticus (Wang et al., 2006). Forty-eight psyllid RNA samples positive for Ca. L. asiaticus were collected for further assay (Fig. S1). Contamination of genomic DNA was removed by RNA treatment with a TURBO-DNA free kit (Ambion, Austin, TX, USA), and the RNA was eluted in 30 μL of water. Checking for DNA contamination in the plant or psyllid RNA samples was performed by a normal PCR test with primers targeting the 16S rDNA of Ca. L. asiaticus (Table S2). No detectable DNA contamination was observed in either plant or psyllid RNA samples (data not shown). RNA quantification was performed using a NanoDrop ND-1000 spectrophotometer (Nanodrop Technologies, Wilmington, DE, USA) and stored at −80 °C.

Selection of Ca. L. asiaticus genes for QRT-PCR analysis

To narrow down the genes, we decided to neglect most housekeeping genes, and focused on putative virulence genes, genes with putative domains and motifs found in virulence factors and genes encoding metabolic pathways, transporters, motility and signal transduction. The list of genes containing a signal peptide was accessed from Integrated Microbial Genomes (http://img.jgi.doe.gov/), predicted using the SignalP v3.0 program (http://www.cbs.dtu.dk/services/SignalP/). The transmembrane domain was predicted by the SMART program (http://smart.embl-heidelberg.de/). The remaining proteins encoded by Ca. L. asiaticus were analysed based on their annotations in the GenBank and KEGG databases.

QRT-PCR analysis

All QRT-PCRs were performed using an Applied Biosystems 7500 Fast Real-time PCR system (Foster City, CA, USA) with a QuantiTect SYBR Green RT-PCR kit (Qiagen). The primers were designed from the sequence of the Ca. L. asiaticus genome using DNASTAR software. The total reaction volume of one-step QRT-PCR was 25 μL and contained 2 × QuantiTect SYBR Green RT-PCR Master Mix (12.5 μL), 10 μm gene-specific primers (1.25 μL), QuantiTect RT Mix (0.5 μL) and 50 ng of RNA template (1 μL). 16S rRNA was used as the endogenous control. Reactions were incubated at 50 °C for 30 min, and at 95 °C for 15 min, cycled (40 times) at 94 °C for 15 s, 54–56 °C for 30 s and 72 °C for 30 s. Melting curve analysis was conducted to verify the specificity of the QRT-PCR products. The products were also run on a 2% agarose gel to confirm the presence of only a single band. Two technical replicates and three biological replicates were used for each of the genes. The relative fold change in target gene expression was calculated using the formula 2^ΔΔCt (Livak and Schmittgen, 2001). Statistical analysis of all data was conducted by Student's t-test (SAS v9.2).

Acknowledgements

This work was supported by the Citrus Research and Development Foundation.