Candidatus Liberibacter: From Movement, Host Responses, to Symptom Development of Citrus Huanglongbing

Candidatus Liberibacter spp. are known for causing the devastating citrus disease huanglongbing (HLB) and zebra chip disease of solanaceous crops. ‘Candidatus Liberibacter asiaticus’ (CLas) and ‘Candidatus Liberibacter americanus’ (CLam), both vectored by Asian citrus psyllid (ACP; Diaphorina citri), and ‘Candidatus Liberibacter africanus’ (CLaf), vectored by African citrus psyllid (Trioza erytreae), cause citrus HLB (Bové 2006; Jagoueix et al. 1994; Teixeira et al. 2005). ‘Candidatus Liberibacter solanacearum’ (CLso; alternate name ‘Ca. L. psyllaurous’), vectored by potato psyllid (Bactericera cockerelli), causes diseases on plants of family Solanaceae, such as potato zebra chip disease (Abad et al. 2009; Munyaneza 2012). ‘Candidatus Liberibacter europaeus’ (CLeu), vectored by pear psyllid (Psylla pyri), infests scotch broom to cause symptoms including stunted growth, shortened internodes, and leaf dwarfing (Thompson et al. 2013). The only member of this group that has been cultured is Liberibacter crescens (Lcr), which was isolated from mountain papaya in Puerto Rico (Leonard et al. 2012).

HLB is a century-old disease and was reported from different regions of the world under various names, including in China as ‘yellow shoot disease’ (HLB) (Reinking 1919), in the Philippines as ‘mottle leaf’ (Lee 1921), in India as ‘dieback’ (Raychaudhuri et al. 1969), in Indonesia as ‘phloem degeneration’ (Aubert 1992), and in South Africa as ‘greening’ (Oberholzer et al. 1965). The Chinese word “Huanglongbing” can be literary translated as “yellow dragon disease (meaning yellow shoot disease in local dialect),” representing the shoots yellowing symptom. Research on HLB gained more attention after its identification in Brazil (2004) (Texeira et al. 2005; Coletta-Filho 2004) and the United States (Halbert 2005; Kumagai et al. 2013). In Florida, 100% of groves with mature citrus trees are HLB positive, which has caused 75% loss of gross citrus production compared with the pre-HLB era (Milne et al. 2018).

All commercial citrus varieties are susceptible to HLB. HLB management in citrus production areas with low HLB incidence relies on the early diagnosis of the disease, replacement of infected trees with healthy plants, and extensive control of ACP population. Region-wide comprehensive implementation of such a strategy successfully controls citrus HLB in a citrus production region of over 110,000 ha of citrus acreage (Yuan et al. 2021). However, this approach is no longer practical in HLB endemic regions with high HLB incidences, such as Florida. The search for resistance or tolerance remains an important way to mitigate or control the impact of the disease. This review summarizes recent progress in studies on Ca. Liberibacter spp. movement as well as -omics analyses of host responses and how Ca. Liberibacter spp. lead to disease development.

BACTERIAL MOTILITY MODES: IS CANDIDATUS LIBERIBACTER MOTILE?

Bacteria are capable of self-propelled movement toward nutrients and suitable environments using appendages such as flagella and pili. Flagella are used for swimming and swarming motility, although this apparatus may also be involved in perception and initial adherence to surface. Swimming motility is a flagella-driven active movement of individual bacterium in liquid to low-viscous medium, while swarming motility is performed by a group of bacteria on a solid or semisolid surface (Kearns 2010). Twitching motility is mediated by type IV pili and involves the translocation of bacteria on solid or semisolid surfaces by attachment, and extension and retraction of type IV pili (Mattick 2002). Gliding movement is an active surface movement occurring along the bacterial long axis that does not require flagella or pili (McBride 2001). Finally, sliding is a passive mode of surface translocation facilitated by surfactants and changes in environment (Hölscher and Kovács 2017).

Ca. Liberibacter spp. genomes encode genes required for the flagella and type IV pili machineries. The expression of those apparatus is upregulated in CLas cells isolated from D. citri, indicating that flagella and type IV pili apparatus may play an important role during CLas colonization and propagation inside psyllids (Andrade and Wang 2019; Andrade et al. 2020; De Francesco et al. 2021). Although low expression of CLas flagella and type IV pilus genes was observed in plant host (Andrade and Wang 2019; Andrade et al. 2020; De Francesco et al. 2021), they may also contribute to bacterial movement or adherence as well as virulence in planta.

Flagellar synthesis in bacteria requires complex regulation that involves perception and transduction of environmental signals. The gene clusters responsible for flagellar synthesis and function include more than 50 genes, which may vary among different bacterial phyla (Liu and Ochman 2007). Flagella in Ca. Liberibacter spp. are encoded by three clusters in the chromosome (Andrade et al. 2020). The flagellar machinery can be divided into eight parts: basal body, motor, switch, hook, filament, cap, junction, and export (Macnab 2003). The basal body consists of a MS ring (FliF), LP ring (FlgH and FlgI), distal rod (FlgG), and proximal rod (FliE, FlgB, FlgC, and FlgF) (Imada 2018; Macnab 2003). Bacterial flagellar hook transmits the torque produced by the basal body to the flagellar filament (Macnab 2003). The motor is made of two parts: stator (MotA, MotB) and rotor (FliG) (Chang et al. 2019; Macnab 2003). The C ring (FliG, FliM, FliN) constitutes the rotor/switch. The hook is a helical structure made of multiple subunits of FlgE. FlgD is a hook capping protein required for flagellar assembly but does not participate in final structure (Liu and Ochman 2007; Macnab 2003; Samatey et al. 2004). The filament is a helical structure comprising approximately 20,000 flagellin subunits (Wang et al. 2017a). The flagellin is encoded by fliC (Macnab 2003). The hook and filament are joined by two junction proteins, FlgK and FlgL (Song et al. 2020). The tip of the filament is made of the annular pentameric cap protein (FliD) (Cho et al. 2017; Postel et al. 2016). The flagellar axial protein components are transported to the central channel of growing flagellum through flagellar type III export apparatus, which consists of FlhA, FlhB, FliH, FliI, FliJ, FliO, FliP, FliQ, and FliR (Fukumura et al. 2017; Imada 2018). Flagellar machinery in rhizobiacea, which includes Ca. Liberibacter, shows some specific characteristics that will be discussed in the following section.

Ca. Liberibacter spp. genomes encode a type IV pili-c or Tad (tight adherence) pilus (Andrade and Wang 2019). The Tad pilus gene cluster in Ca. Liberibacter encodes a complete set of proteins that are necessary for assembly and function of the Tad pilus machinery, which includes secretion platform formed by CpaABCDEF and TadCD, secretin CpaI, pseudopilins TadEF, membrane anchor for pilins TadG, and pilins Flp1-to-Flp7 (Andrade and Wang 2019). It was reported that Tad pili can promote surface colonization, cell-to-cell aggregation, and biofilm formation and contribute to bacterial virulence of different pathogens (Sangermani et al. 2019). In addition, Tad pili are retractable and might contribute to bacterial motility (Ellison et al. 2017, 2019). The presence of flagellar apparatus and pili in Ca. Liberibacter spp. suggests active motility. However, the pattern of expression and contribution of motility in colonization and movement inside insect and plant hosts and virulence in planta need further investigation.

Flagellar synthesis and assembly genes in Ca. Liberibacter spp.

Analysis and comparison of multiple Ca. Liberibacter genomes show that 35 genes associated with flagellar biosynthesis and assembly are conserved in CLas, Lcr, CLam, CLso, and CLaf (Fig. 1). Apparently, a near complete set of flagellar biosynthesis and assembly genes, including the genes required to encode basal body, hook, motor, junction protein, filament, and flagellar export apparatus, are conserved among Ca. Liberibacter spp. (Fig. 1) (Andrade et al. 2020). Notably, members of Rhizobiaceae, including Ca. Liberibacter species, lack FliJ (a cytoplasmic chaperon), FliH/FliO (components of export apparatus), and FliD (filament cap) (Andrade et al. 2020). In Ca. Liberibacter spp., the flagellin is encoded by flaA (Andrade et al. 2020). The MotD protein of Sinorhizobium meliloti and related α-proteobacteria acts as a regulator of flagellar hook length and the function of MotD is performed by FliK (Eggenhofer et al. 2006). The FlgJ of β-proteobacteria and γ-proteobacteria consists of canonical dual domain: a N-terminal scaffold for rod assembly and a C-terminal peptidoglycan hydrolase (PGase) (Nambu et al. 2006; Herlihey et al. 2014). FlgJ contains only a single rod assembly domain in α-proteobacteria and utilizes a PleA (a PGase) homolog for successful flagellar assembly (Viollier and Shapiro 2003; Nambu et al. 2006). FlgJ of Ca. Liberibacter spp. also consists of a single rod assembly domain but lacks the peptidoglycan hydrolase domain, which is involved in the elongating flagellar rod penetrating through the peptidoglycan (PG) layer. This seems to be complimented by a conserved peptidoglycan hydrolase encoded by Ca. Liberibacter spp., which contains a periplasmic signal peptide and a lytic transglycosylase (SLT) domain (Andrade et al. 2020).

Download as PowerPoint

Type IV Tad pili synthesis and assembly genes in Ca. Liberibacter spp.

The Tad pilus in bacteria was acquired after interkingdom transfer from archaea and resembles ‘EppA-dependent’ (Epd) pilus (Denise et al. 2019). Tad pili mediate the surface attachment of planktonic bacteria, aiding in surface colonization and biofilm formation (Pu and Rowe-Magnus 2018; Sangermani et al. 2019). Tad pili undergo repeated cycles of extension and retraction after attachment, which promotes walking-like surface motility (Ellison et al. 2019; Sangermani et al. 2019; Pu and Rowe-Magnus 2018). Tad pili play vital roles in host colonization and promote virulence in several pathogenic bacteria (Motherway et al. 2011; Nykyri et al. 2013; Pu and Rowe-Magnus 2018; Wairuri et al. 2012). The Tad pili encoding gene clusters are conserved among α-proteobacteria, including Ca. Liberibacter spp. (Fig. 2) (Andrade and Wang 2019; Duan et al. 2009; Lin et al. 2011; Mignolet et al. 2018; Naranjo et al. 2019). The Tad genes in Ca. Liberibacter spp. encode all the components required for assembly and function of Tad pili (Andrade and Wang 2019). Approximately 15 to 20 genes encode Flp pilins, pseudopilins (TadE/CpaJ and TadF/CpaK), pilotin (TadD/CpaO), prepilin peptidase (CpaA), secretin (CpaC), inner membrane core (TadB/CpaG and TadC/CpaH), periplasmatic subunits (CpaB and CpaD), CpaI, TadG, and ATPase (CpaF), and they are conserved among Ca. Liberibacter spp., and in closely related members of Rhizobiaceae, such as Agrobacterium tumefaciens, and S. meliloti (Fig. 2) (Andrade and Wang 2019). Comparative gene expression analysis of the genes encoding the Tad pili apparatus of CLas cells isolated from both citrus and psyllid showed that this system is upregulated in psyllids (Andrade and Wang 2019). Interestingly, a CLas Flp pilin named Flp3 was found to be highly expressed in psyllids and might be involved in biofilm formation during colonization of the insect vector (Andrade and Wang 2019). Likewise, a recent study of the transcriptome profiling of CLas in citrus and psyllids shows the downregulation of Tad genes in plant host (De Francesco et al. 2021). This same study reported the upregulation of the pilin Flp1 in plant host, and validated the upregulation of Flp3 in psyllid, as previously demonstrated by Andrade and Wang (2019). To confirm whether Flp1 and Flp3 play a role in the colonization of CLas in citrus plant and psyllids, respectively, further investigations are required.

Download as PowerPoint

Expression of motility appendages in Ca. Liberibacter spp.

TEM (transmission electron microscopy) micrographs of Lcr grown in vitro revealed the presence of multiple flagella (Andrade et al. 2020). The negatively stained TEM micrographs of CLso prepared from the midgut of potato psyllids showed both flagella and pili like appendages (Cicero et al. 2016). However, expression of appendages only occurs in certain environments, as CLso from biofilm or dissected digestive organs of potato psyllids lacked pili or flagella (Cicero et al. 2016). The diameter of pili-like structures ranged from 0.005 to 0.008 µm and were observed in bundles of 10 or more. Flagella-like appendages of 0.02 µm diameter were also observed on CLso cells prepared from the midgut of potato psyllids (Cicero et al. 2016). TEM micrographs of few CLas cells isolated from the midgut of ACPs exhibit thread-like appendages. However, CLas cells from grapefruit (Citrus paradisi) seed coats and parasitic dodder stem did not show any external appendages (Andrade et al. 2020). It remains to be determined whether this results from loss of flagella during sample preparation or nonexpression in certain compartments or stages of infection. TEM micrographs of dissected plant tissues (seed coat, root, and young flush) from CLas-infected Duncan grapefruit and Valencia sweet orange demonstrated the attachment of CLas cells to phloem cell plasma membrane through unknown filamentous structures, which may be formed by CLas appendages or fusion of the cell membranes from CLas and plant host (Achor et al. 2020).

Transcriptome analysis of CLso in the gut of potato psyllids revealed that the pili assembly and flagellin encoding genes are among the most expressed genes (Yao et al. 2016). Similarly, analysis of CLas living in the gut of ACPs found that pilins and flagella synthesis and assembly genes are among the most highly expressed genes (Darolt et al. 2021). Quantitative RT-PCR expression analysis suggests the enhanced expression of Ca. Liberibacter spp. flagellar and pilin-associated genes in psyllids compared with the plant host (Andrade et al. 2020; Yan et al. 2013), which was corroborated by transcriptome analysis (De Francesco et al. 2021). It is yet to determine how pili and flagellar genes are regulated in Ca. Liberibacter and their roles in colonization of the insect vector and the plant host. The export apparatus of flagellar and type IV pili assembly systems possesses the specialized secretion system for flagellin and pilins, respectively. However, the flagellar export apparatus has been shown to be involved in the secretion of virulence factors in the human pathogen Yersinia enterocolitica (Young et al. 1999). Ca. Liberibacter spp. contain the basic Sec-dependent secretion system and type I secretion system, but do not encode other secretion systems, such as type II, III, and IV, to deliver the virulence factors into host cells. It was suggested that CLas encodes two autotransporters LasAI and LasAII. However, no signal sequence was identified in LasAI or LasAII (Hao et al. 2013). Thus, it is legitimate to inquire whether the motility-related export machinery participates in the secretion of Ca. Liberibacter virulence factors.

Ca. Liberibacter spp. movement inside the insect vector.

Although psyllids can acquire Ca. Liberibacter at any stage of life, nymphal stages are more efficient in acquisition (Pelz-Stelinski et al. 2010). However, adult ACPs inoculate CLas more efficiently than nymphs when CLas are acquired by early instar nymphs (Ammar et al. 2020). Unlike Xylella fastidiosa, CLas and CLso are both carried by their insect vectors in a persistent circulative propagative manner. CLas and CLso can circulate through different parts of insect beyond the stylet and foregut, are passed through the molting process, and can multiply inside psyllids (Ammar et al. 2016; Fisher et al. 2014; Perilla-Henao and Casteel 2016; Tang et al. 2020). It was reported that CLas enters gut cells by endocytosis and form Liberibacter containing vacuoles, which is followed by the formation of endoplasmic reticulum-related and replication permissive vacuoles. The authors conducted immunolocalization studies and demonstrated that CLas employs endo/exocytosis-like mechanisms for invasion and egress (Lin et al. 2011).

Microscopy-based detection showed CLas colonization in the midgut, malpighian tubules, hemocoel, filter chamber, salivary glands, ovaries, vacuoles of the endoplasmic reticulum, muscle, and fat tissues of ACPs (Ammar et al. 2011, 2016, 2019; Ghanim et al. 2016; Hall et al. 2013). How Ca. Liberibacter spp. move to all these different body parts remains unknown. Apparently, Ca. Liberibacter spp. produce external movement appendages inside the insect vector, but a comprehensive understanding of the mode of movement remains elusive. The expression and maintenance of flagella in insect vector suggests the importance of flagella-mediated movement. Initial attachment and sense of different surfaces is probably mediated by flagella, whereas development of bacterial biofilm is promoted by pili. The colonization of different ACP organs indicates that attachment structures like the Tad pili may be important for colony formation and in-colony movement within the insect host.

Ca. Liberibacter spp. movement inside plant hosts.

Understanding the movement of Ca. Liberibacter spp. in plants could be crucial to develop suitable control strategies against HLB. Microscopy-based observations suggest that CLas is mostly confined to citrus phloem sieve elements (Folimonova and Achor 2010; Hilf 2011). However, CLas was observed in nucleated nonsieve element cells although the identity of these CLas-containing nucleated cells (whether they are developing sieve elements, companion cells, or phloem parenchyma) remains unknown (Achor et al. 2020). Similarly, CLso colonizes in the sieve elements (Secor et al. 2009; Teresani et al. 2014). CLso can move from transmission site to stem within a week after inoculation by psyllid feeding on tomato and potato plants (Levy et al. 2011). Levy et al. (2011) reported that CLso was most frequently detected in the upper tier and middle tier of leaves within 2 to 3 weeks after infestation. CLso also moves to newly emerging leaves and other sink tissues like stolon and root (Pitman et al. 2011; Wen et al. 2009).

Insect vectors inoculate Ca. Liberibacter spp. in plant host through feeding. Young leaves of citrus trees receive the first CLas inoculum as those tissues are the preferred feeding and reproduction sites of psyllids (Bové 2006; Pandey et al. 2021; Pluke et al. 2008; Sétamou et al. 2016; Wenninger and Hall 2007). Apparently, CLas cells remain in the young flush until the leaves are mature enough to turn into a source (Andrade et al. 2020; Pandey et al. 2021). The initial colonization of CLas at feeding site is useful for early detection (Pandey and Wang 2019).

Ca. Liberibacter spp. movement occurs primarily vertically along the sieve tubes while traversing the sieve pores (Achor et al. 2020; Folimonova and Achor 2010; Lopes et al. 2009; Nissinen et al. 2014; Wang et al. 2017b). Sieve pore diameters of citrus range from 120 to 418 nm and are traversable for CLas cells, whose diameters range from 100 to 300 nm inside citrus phloem (Bové 2006; Hilf 2011; Koh et al. 2012). CLas-induced callose plugging greatly reduces the diameters of sieve pores (Folimonova and Achor 2010; Kim et al. 2009). However, this plugging is ineffective at restricting CLas movement, as the flexible and pleomorphic shape of CLas cells allow them to pass through narrowed sieve pores (Achor et al. 2020; Ammar et al. 2019).

CLas movement is predominantly passive by floating freely with the phloem sap from sources (e.g., mature leaves and roots) to sinks (e.g., developing roots, young leaves, flushes, and fruit) (Achor et al. 2020; Johnson et al. 2014; Kim et al. 2009; Raiol-Junior et al. 2021; Wang et al. 2017b). However, the rates of CLas movement (between 2.9 and 3.8 cm·day⁻¹) occur at a much slower pace than the flow of citrus phloem sap (0.34 m·day⁻¹) (Etxeberria et al. 2016; Raiol Junior et al. 2021). Systemic spread of CLas inside citrus trees takes months to years. CLas begins in young flush after transmission from psyllids, then move to the roots, followed by young leaves that emerge post-ACP feeding, and finally to mature leaves (Pandey et al. 2021). The fact that CLas does move inside mature leaves against the flow of phloem sap suggests that active movement may occur inside citrus trees. Yet evidence does not support the expression of flagella-like external appendages on CLas cells inside the plant hosts (Andrade et al. 2020; Bové 2006) (De Francesco et al. 2021). Pathogenic bacteria may avoid flagellar production to escape the recognition and consequent host defense response (Chatnaparat et al. 2016; Ramos et al. 2004; Yu et al. 2013). However, TEM micrographs suggest that CLas cells can attach to the plasma membrane through unknown external filaments or pathogen-host plasm membrane cell fusion inside the phloem and change the body shape to traverse through the sieve pores (Achor et al. 2020). Although genes encoding Tad pili and flagella machineries are downregulated in plant, it cannot be ruled out that basal expression of Tad pili-like external appendages, which participate of surface attachment and twitching motility, and/or temporary expression of flagella, or gliding motility, may play a role in CLas active movement.

THE CA. LIBERIBACTER−PLANT HOST INTERFACE: BACTERIAL ADAPTATION AND HOST DEFENSE RESPONSE

Plant pathogens manipulate the global expression, metabolism, and host defense responses and alter the fitness in both the insect vector and the plant hosts (Chatterjee et al. 2008; Eigenbrode et al. 2018; Killiny 2022). CLas infection modifies the physiology, reproduction, dispersal, movement, and behavior of ACPs to spread to a greater extent (Molki et al. 2019; Pelz-Stelinski and Killiny 2016; Ren et al. 2016; Vyas et al. 2015). The Ca. Liberibacter and insect vector interface was recently reviewed (Galdeano et al. 2020; Killiny 2022). Here, we reviewed the recent progress in Ca. Liberibacter-plant host interface and the host responses.

GLOBAL GENE EXPRESSION OF THE PLANT HOST

Global transcriptomic and proteomic profiling of host plants with mild to severe Ca. Liberibacter spp. infections have been extensively studied and reviewed (Chin et al. 2021; da Graça et al. 2016; Franco et al. 2020; Huot et al. 2018; Martinelli et al. 2012, 2016; Rawat et al. 2015; Wang and Trivedi 2013; Wang et al. 2016; Wei et al. 2021; ). Comparative transcriptomic analysis of tolerant and susceptible varieties has been instrumental for understanding host responses to CLas infection (Rawat et al. 2015; Wang et al. 2016; Yu et al. 2017). Comparative transcriptome analysis of the HLB-tolerant “Jackson” grapefruit and the HLB susceptible “Marsh” grapefruit trees shows reduced expression of DMR6 (putative 2OG-Fe(II) oxygenase)-like genes, beta glucanases, expansins, and DET2 in the tolerant variety, which are likely involved in the suppression of host immunity (Wang et al. 2016). However, several susceptible varieties exhibit induced expression of defense-related genes and higher expression of callose synthase and cell wall degradation enzyme-encoding genes (Fan et al. 2012; Fu et al. 2016; Hu et al. 2017; Kim et al. 2009; Koh et al. 2012). The tolerant varieties show induced expression of NPR1, BS-LRR, RLK, cellulases, expansins, pectinesterases, and host defense and pathogenesis-related (PR) genes, but relatively lower effects on starch and photosynthesis (Arce-Leal et al. 2020; Curtolo et al. 2020; Fan et al. 2012; Hu et al. 2017; Mafra et al. 2013; Wang et al. 2016; Yu et al. 2017). The midrib tissues from the leaves of “Shatangju” mandarin (Citrus reticulata Blanco “Shatangju”) exhibit upregulation of virulence, stress response, transport system, flagellar assembly, lipopolysaccharide biosynthesis, and cell surface structure associated genes compare with the fruit piths (Fang et al. 2021).

The host responses may differ based on the age of leaf tissues and the stage of bacterial infection (Ribeiro et al. 2021; Wei et al. 2021). The constitutive and induced gene expression of immunity-related genes in mature leaves of citrus may provide greater resistance or tolerance against bacterial infection (Ribeiro et al. 2021). Wei et al. (2021) demonstrated that the pathways related to signaling, transcription factors, defense, hormone, and photosynthesis in Valencia sweet orange are differentially expressed as early as 1 day post-ACP-mediated inoculation of CLas (Wei et al. 2021). They further showed a burst of DEGs (differentially expressed genes) associated with defense, photosynthesis, secondary metabolites, and ATP biosynthetic and glycolytic pathways at 5-day postinoculation (Wei et al. 2021). Relatively HLB tolerant lemon shows a lesser effect on photosynthetic pathways, induced expression of protease inhibitors at presymptomatic stage compared with susceptible sweet orange (Chin et al. 2021; Ramsey et al. 2020). They also had a greater accumulation of micronutrients. RNA-sequencing data showed that CLso infection resulted in differential expression of 397 and 1,027 genes in two different varieties of potato, Atlantic (more susceptible) and Waneta, respectively (Levy et al. 2017). Levy et al. (2017) showed that the majority of the DEGs were downregulated in both varieties, including the genes involved in photosynthesis, primary/secondary metabolism, signaling, and stress (Levy et al. 2017). The reprogramming of host global expression in response to other Ca. Liberibacter spp. needs more robust investigation.

Ca. Liberibacter infestation causes substantial interference with the proteome of host plants (Franco et al. 2020; Nwugo et al. 2016; Ramsey et al. 2020; Yao et al. 2019). Nwugo et al. (2016) measured the differential expression of proteins in grapefruit leaves and found them to be greatly influenced by heat treatment. Glucosidase II beta subunit-like protein, chlorophyll a/b binding protein, a glutathione S-transferase, a ferritin-like protein, and a putative lipoxygenase protein were all downregulated during CLas infection but upregulated in the presence of heat treatment (Nwugo et al. 2016). A comparative proteome analysis of Navel orange (a highly susceptible variety) and Volkamer lemon (Citrus volkameriana) (a moderately tolerant variety) shows the upregulation of four glutathione-S-transferases, which are involved in radical ion detoxification, in Volkamer lemon (Martinelli et al. 2016). Symptomatic fruit of Valencia sweet orange trees grafted on Swingle citrumelo (C. paradisi × Poncirus trifoliata) rootstock showed the degradation of proteins involved in glycolysis, the tricarboxylic acid (TCA) cycle, and amino acid biosynthesis compared with the healthy fruit (Yao et al. 2019). Phloem proteins of the HLB-susceptible sweet orange variety, Washington navel 10-month postgraft inoculation of CLas display suppression of plant metabolism and translation but induced expression of defense-related proteins such as peroxidases, proteases, and protease inhibitors (Franco et al. 2020). A proteomic comparison of asymptomatic and symptomatic stages of Newhall sweet orange plants with high CLas titers compared with their healthy counterpart shows the suppression of the pathway involved in the photosystem I and II light reactions throughout the infection process (Li et al. 2021). Li et al. (2021) further showed that starch biosynthesis was upregulated during the asymptomatic stage. Host defense responses were induced in leaf petioles of symptomatic leaves and the salicylic and jasmonic acid content exhibited positive correlation with phytohormone biosynthesis-related proteins in response to CLas infection.

METABOLISM OF PLANT HOST

Metabolomics is a powerful tool for studying plant behaviors owing to its breadth and its ability to look directly at relevant phenotypes (Pandita et al. 2021). By detailing the composition of metabolites in a sample, researchers can directly observe many key traits, including processes like nutrition, signaling, and defense that are central to host–pathogen interactions. Many metabolomic studies have been conducted on HLB–citrus interactions. However, the breadth and specificity of metabolomics come with the drawback of enormous variation. Metabolomic responses can vary widely based on host variety, tissue type, infection status, plant age, and other variables on the host or pathogen side. However, some general trends in citrus metabolomic response to HLB have emerged.

Sugars.

Sugars are one of the most common plant metabolites, are sources of energy for both eukaryotes and prokaryotes, and are the cargo transported in the phloem (Dinant et al. 2010). HLB blocks phloem flow and changes the expression of sugar metabolic genes, which can alter sugar levels in infected tissue (Kim et al. 2009). While CLas encodes a glucose/galactose transporter and likely relies on host-derived sugars for energy, sugars can also play a role in plant defense responses (Duan et al. 2009; Rojas et al. 2014; Wang and Trivedi 2013). Therefore, both the host and CLas likely influence sugar levels in infected tissues to control the phloem environment. Sugar levels might be influenced by the blockage of phloem flow, the incorporation of sugars into starches, sugar uptake by CLas for nutrition, or reduced photosynthesis of the plant. Overall, recent work suggests a general decline in sugar levels in host tissue during infection. Albrecht et al. (2016) conducted two untargeted metabolomic studies of both susceptible and tolerant citrus varieties. In one, several sugars had increased levels upon infection at 8 months postinoculation, but levels of the same sugars were lower 2 months later. In the tolerant variety, glucose, fructose, and raffinose were higher in the infected samples at both time points. In their second study, levels of several sugars were commonly lower in infected tissue. A similar dynamic response was found in a study of Valencia and Hamlin sweet orange fruit (Chin et al. 2014). In asymptomatic Hamlin fruit, fructose, glucose, and sucrose had increased concentrations, but levels in symptomatic fruit of both varieties were reduced. Hung and Wang (2018) examined targeted metabolites in Hamlin sweet orange, mandarins, and grapefruits in response to infection. Most sugars were unchanged, except for fructose being lowered in mandarin and grapefruit, glucose and galactose lowered in grapefruit, and melibiose up in Hamlin and mandarin. A reduction extends to juice, as multiple studies have found reduced sugar concentrations in the juice of infected fruit (Slisz et al. 2012; Yao et al. 2019). The variations observed in sugar responses underscore the complexity of the CLas–citrus interaction. While the most common response in the above studies is a reduction of sugar levels, more work is needed to dissect the most important drivers of changes in sugar levels and how any changes affect CLas pathogenesis.

Amino acids.

Besides acting as important nutrients and sources of nitrogen, amino acids are the precursors for key plant defense compounds and hormones. Several studies have found increases in the levels of amino acids during infection (Albrecht et al. 2016; Hung and Wang 2018). In a metabolic study of leaf tissue of multiple susceptible and tolerant citrus varieties, Albrecht et al. (2016) found that proline was the most commonly increased compound during CLas infection. This was also true in grapefruit in a separate targeted metabolomic study of several citrus genotypes (Hung and Wang 2018). This may represent an immune response by the plant, as proline commonly accumulates in plants in response to stress and may play a role in cell wall development and defensive signaling (Hayat et al. 2012; Kishor et al. 2015). What is not clear from these studies is if proline accumulation is required for tolerance or if increasing proline levels in susceptible varieties could improve their tolerance to HLB. Furthermore, this effect may be tissue-specific, as a metabolic study of fruit juice found a reduction in proline levels in the juice of infected fruit (Slisz et al. 2012). A similar pattern is seen for arginine, which can also contribute to immune response through the production of NO (Winter et al. 2015). Arginine was found to be increased in the leaf tissue of infected grapefruit and Cleopatra mandarins but was found to be reduced in infected juice (Albrecht et al. 2016; Hung and Wang 2018; Slisz et al. 2012).

Phenylalanine, a precursor for many stress-response and defensive compounds, is also commonly increased upon infection (Chin et al. 2014; Hung and Wang 2018; Slisz et al. 2012). Phenolic compounds that may be protective against invading pathogens and derived from phenylalanine were shown to be increased in infected citrus leaves (Hijaz et al. 2013).

Fatty acids.

A recent study examined the fatty acid content of sweet orange leaves and found a clear separation between infected and healthy tissues (Suh et al. 2018). Specifically, Suh et al. (2018) found that symptomatic tissues had lower levels of four long-chain fatty acids. The levels of these fatty acids could be partially restored by thermotherapy, strengthening the link between HLB and their reduction. Fatty acids can play a role in plant immune response, both by interacting with signaling molecules and by participating in the production of reactive oxygen species (Yaeno et al. 2004). It remains unclear if the reduction in long chain fatty acids observed by Suh et al. (2018) represents a consequence of host immune response or if they are actively being degraded by the pathogen. Further study of the dynamics of fatty acid levels in response to HLB or the response in other citrus varieties may be a fruitful avenue of future research.

Metabolic differences between susceptible and tolerant varieties.

HLB tolerant citrus varieties are a potential source of anti-HLB compounds and other defensive strategies. Many studies have examined metabolic differences between tolerant and susceptible varieties to find signatures of tolerance. Some of these tried to find compounds that might directly attack CLas or its vector. Killiny et al. (2018) measured the metabolome of three HLB-tolerant varieties including finger lime (Microcitrus australasica (F. Muell.)) (Killiny et al. 2018). Among the compounds that were present at higher concentrations in the tolerant varieties were several volatile organic compounds (VOCs) that have direct antimicrobial effects, including citronellal, farnesol, and others. They also found high levels of amino acids like phenylalanine and the phenolic precursor shikimic acid in finger lime. However, two other tolerant cultivars, LB8-9 and the potentially tolerant Bingo mandarin, had differing profiles of VOCs, suggesting that tolerance can arise from multiple mechanisms. For example, the antimicrobial VOC thymol was high in LB8-9 but almost absent in mandarin hybrid Bingo.

Other metabolomic studies of tolerance have focused on basic nutrients. Albrecht et al. (2016) compared a susceptible Cleopatra mandarin with several tolerant varieties. They found that the tolerant and moderately tolerant varieties could be statistically separated from the susceptible varieties, suggesting that there might be a metabolic signature of tolerance. The metabolites that were significantly altered in tolerant varieties were primarily unknown compounds, but several sugars were significantly reduced. Killiny (2017) examined the metabolomes of 14 varieties to find a signature of tolerance. Unlike in Albrecht et al. (2016), they did not find a clear global separation between susceptible and tolerant, and they did not find that sugar concentrations correlated to tolerance. However, methodological differences make direct comparisons between the two studies difficult, as Killiny et al. (2018) examined phloem sap rather than leaf tissue, used a different extraction method, focused on known metabolites, and used different varieties. Although they found no clear differences among sugars, they did find that phenylalanine, tryptophan, and tyrosine, all precursors to defensive compounds, correlated positively with tolerance, as did some organic acids like shikimic acid, GABA, and fumaric acid (Albrecht et al. 2016). These results mirrored a previous study by the same group using the same techniques, which examined the metabolomes of the phloem of 13 citrus varieties and found that tolerant varieties have higher levels of phenylalanine, tryptophan, and tyrosine (Killiny and Hijaz 2016). In this same study, a few compounds were associated with susceptibility, including some sugar acids and the phenolic amine synephrine. Synephrine has been used as a dietary supplement in humans and is commonly extracted from bitter orange, but its role in plants is unknown (Stohs et al. 2020). It remains unclear if these compounds are good nutrient sources for CLas or if they are indicative of lowered defense. A recent targeted analysis of 178 metabolites in susceptible Murcott mandarin and tolerant LB8-9 mandarin found a clear separation between the tolerant and susceptible samples in both healthy and infected leaf tissue (Suh et al. 2021). Pathways associated with glutamine and asparagine metabolism, purine metabolism, and the synthesis of auxins were upregulated in both healthy and infected LB8-9. Curiously, salicylic acid (SA)-related metabolites were induced in the susceptible variety but not in the tolerant variety, suggesting that LB8-9 tolerance may involve reducing some defense responses in favor of growth. This is different from the results by Albrecht et al. (2016), who compared susceptible Cleopatra mandarin to two tolerant hybrids and found the SA precursor salicin in higher levels in the tolerant varieties. However, salicin was not directly measured by Suh et al. (2021).

Taken together, these studies indicate that the overall metabolome seems to depend more on the variety rather than the susceptibility to HLB. Some specific metabolites associated with basal defense, either through direct antimicrobial effects or as precursor for defensive signaling compounds, may help explain tolerance to HLB. However, tolerance may be attained by different varieties through multiple pathways. Understanding and using multiple pathways to combat HLB may help to provide more durable, lasting control strategies.

CLAS ADAPTATIONS TO LIFE IN ITS HOST

Upon entry into the host phloem, CLas must rapidly alter its behavior to an environment that is vastly different from its insect vector. Despite its reduced genome, CLas encodes a variety of traits that allow it to grow, spread, and evade host defenses.

Nutrient acquisition.

Although the inability to culture or genetically modify CLas limits our understanding of its nutrient requirements, we can use genomics to predict which metabolic pathways are important in planta. As intracellular endosymbionts, they have lost many metabolic traits (Duan et al. 2009). Among these missing genes are several parts of the electron transport chain, implying that CLas cannot complete oxidative phosphorylation on its own. However, it does have an ATP/ADP translocase, suggesting that it could obtain ATP from its host. CLas metabolizes a limited set of sugars, such as glucose, fructose, and xylulose but not mannose, galactose, rhamnose, or cellulose. While it can metabolize many amino acids, it only has synthesis genes for phenylalanine, aspartic acid, and lysine, but probably imports the others (Duan et al. 2009).

In addition to scavenging nutrients already present in host phloem, CLas may alter nutrients in its direct environment. For example, ATP, which CLas probably imports for its own use, was found to be increased in infected citrus leaves, likely because of the upregulation of ATP synthase (Pitino et al. 2017). Sugars are also common targets for bacterial pathogens, and CLas infection induces alterations in sugar metabolism in citrus (Albrecht et al. 2016; Cohn et al. 2014; Hung and Wang 2018). However, given that carbon sources are abundant in the phloem, it is not clear if these alterations would significantly affect CLas metabolism, or if their primary effects are on host defense.

Transcriptional control in the phloem.

Pathogens must tightly control their gene expression during infection to efficiently deploy their virulence factors. CLas is no exception. Yan et al. (2013) examined expression changes of a set of 381 predicted virulence-related genes in psyllids versus in planta. They found that 182 of these genes had increased expression in planta and included several transporters, secretion-related genes, flagellar assembly genes, and metabolic genes. How does CLas control these changes in expression? CLas encodes a relatively small number of transcription factors (Duan et al. 2009). Without the ability to culture or genetically modify the bacterium, studying the effects or triggers of these transcription factors is difficult. Recent work suggests the high osmolarity of the phloem may help act as a trigger to transition the bacteria from an insect mode to a plant mode. Pagliai et al. (2014) used two relatives of CLas (Lcr and S. meliloti) to study the activity of the transcription factor LdtR. Inactivation of LdtR in either Lcr or S. meliloti leads to an increased susceptibility to osmotic stress. They found that LdtR is a transcriptional activator that may control cell wall remodeling and response to osmotic stress, in part by the activation of LotP (Loto et al. 2017). After chemically inhibiting LdtR in Lcr, alterations were found in 180 genes, many of which are involved in motility, cell wall metabolism, and energy production (Pagliai et al. 2017). Among the genes apparently activated by LdtR is FerR, a ferredoxin-like regulator whose expression is also increased in Lcr by high osmolarity (Pan et al. 2019). Together, these results suggest that sensing osmotic pressure may help CLas transition between stages in its life cycle. More work could find the mechanism of this sensing and find other triggers that coordinate changes in the CLas life cycle.

EVASION OF HOST IMMUNE RESPONSE

None of the commercial citrus varieties are resistant to HLB, suggesting that CLas might have adapted to citrus relatively recently (Bové 2006). There have been limited chances to develop an effective R gene to provide ETI (Effector-Triggered Immunity) (da Graça et al. 2016). The first layer of plant immune system, PTI (PAMP-triggered immunity), however, probably plays a role in defense against HLB. CLas encodes multiple PAMPs (pathogen-associated molecular patterns) that could be recognized to induce a PTI response (Duan et al. 2009). Evidence suggests that plants can recognize CLas and CLso PAMPs like Flg22 (Hao et al. 2014; Zou et al. 2012). Therefore, to succeed in planta, Ca. Liberibacter spp. must encode virulence factors that counteract PTI. Among these factors are SEC-dependent effectors, which may help to directly counteract PTI. Components of the SEC system are upregulated in planta compared with in psyllids, along with 26 predicted SEC secreted proteins (Yan et al. 2013). Approximately 166 proteins have predicted SEC signals, 86 of which were validated to be exported to the periplasm in Escherichia coli (Prasad et al. 2016). Some of these are likely bacterial factors (e.g., flagellar components, transporters, cell division proteins), but others may be host-targeted effectors.

The role of one of these Sec-Delivered Effectors (SDEs), SDE1, has been characterized in planta. SDE1 was shown to interact with and inhibit several defense-inducible host papain-like cysteine proteases (Clark et al. 2018). When expressed in Arabidopsis, SDE1 promotes the colonization and growth of Pseudomonas syringae, supporting a role in suppressing general host immunity. Transgenic citrus plants expressing SDE1 had increased susceptibility to HLB and increased expression of senescence-related genes (Clark et al. 2020). Besides inhibition of host cysteine proteases, SDE1 was shown to interact with an RNA helicase and induce chlorosis in tobacco (Zhou et al. 2020). Another study expressed SDE1 in tobacco and found that it localized to the chloroplast and induced cell death and callose deposition (Pitino et al. 2016). However, such effects were not observed when overexpressed in citrus (Clark et al. 2020).

The prophage encoded peroxidase (SC2_gp095) of CLas was shown to be a secreted effector, which suppresses the host defense response by controlling the expression of RbohB (respiratory burst oxidase homolog B; involved in H₂O₂-mediated defense signaling in plants) (Jain et al. 2015). SDE15 from CLas interacts with CsACD2 of citrus host and suppresses the plant immunity (Pang et al. 2020). Pang et al. (2020) proposed CsACD2 as potential target for genome editing to generate HLB-resistant plants. The Sec-translocon–dependent effector Lso-HPE1(‘Ca. L. solanacearum’–hypothetical protein effector 1) from CLso suppresses the host immunity and plant cell death (Levy et al. 2020). How effector-mediated manipulation of host defense responses and correlates with the development of disease symptoms remains mostly unexplored in the Ca. Liberibacter spp. pathosystem.

SA is important to citrus defense against HLB. SA levels have been found to be elevated in Valencia orange leaves infected with HLB (Lu et al. 2013) and in sweet orange cultivar Newhall symptomatic and presymptomatic leaves (Li et al. 2021). Salicin, a precursor to salicylic acid, was reduced in susceptible compared with tolerant varieties (Albrecht et al. 2016). Artificially altering SA levels can also affect HLB. Injection of SA into the trunks of Hamlin sweet orange trees reduced both disease incidence and CLas titers in the field (Hu et al. 2017). In addition, overexpression of an inducer of SA-mediated systemic acquired resistance (SAR) increased the tolerance of sweet orange trees to HLB (Dutt et al. 2015).

Given the importance of SA signaling in defense, it is unsurprising that SA is targeted by CLas during infection. CLas encodes the salicylic hydroxylase SahA that can degrade SA (Li et al. 2017; Wang and Trivedi 2013). sahA is upregulated in plants compared with psyllids (Li et al. 2017). Some of the proteases inhibited by the sec-dependent effector SDE1 accumulate in response to SA, suggesting that part of SDE1’s role may be to mute SA responses (Clark et al. 2018). Understanding the induction of SAR in response to CLas and if tolerant varieties can overcome CLas-mediated inhibition of SA and SAR may help us understand how to promote tolerance in new citrus varieties.

DISEASE DEVELOPMENT AND SYMPTOM APPEARANCE AFTER CA. LIBERIBACTER SPP. INFECTION

Disease symptoms.

CLas, CLaf, and CLam cause similar HLB symptoms. However, CLas causes the most severe symptoms of the three. HLB symptoms include blotchy mottled leaves with green islands; short, hardened, and upright leaves; corky veins; premature defoliation, stunted growth; off-season flowering; early fruit drops; small lopsided fruit with poor coloration and aborted seeds; root decline; thinning of canopy; and dieback of twigs and trees (Fig. 3) (Bové 2006; Gottwald et al. 2007; Johnson et al. 2014; Wang 2019; Wang et al. 2017b; Wang and Trivedi 2013). The Zebra chip disease of solanaceous crops, caused by CLso, exhibits leaf curling and scorching; chlorosis; shortened and thickened internodes; enlarged nodes; browning of vascular tissue; aerial tubers, and early necrosis and death (Buchman et al. 2012; Liefting et al. 2009; Munyaneza 2012; Munyaneza et al. 2009). CLeu infests scotch broom to cause the symptoms like leaf dwarfing and stunted growth but has also been reported from pear, apple, and hawthorn as endophytes without any disease symptoms (Raddadi et al. 2011; Thompson et al. 2013).

Download as PowerPoint

Systemic spread of Ca. Liberibacter spp.

Citrus trees may remain asymptomatic for months after CLas transmission by ACPs feeding (Bové 2006; Lee et al. 2015; Pandey et al. 2021; Wang et al. 2017b). However, asymptomatic citrus trees may act as sources of inoculum for CLas acquisition by ACPs and contribute to the disease spread (Lee et al. 2015). Ca. Liberibacter spp. are cryptic in nature and bacterial titers are inconsistent with the tissue-specific symptom appearance (Wang et al. 2017b). Young shoots play a central role in the HLB pathosystem, as they are the preferential feeding sites and exclusive reproduction sites for ACP (Bové 2006; Pluke et al. 2008; Sétamou et al. 2016; Wenninger and Hall 2007). ACP feeding transmits CLas to young flush (Pandey et al. 2021). The young flush works as sink at the early stage, importing nutrients from other tissues. Once the young flushes are mature enough to work as a source, the sieve elements are loaded with sugars to transport to other sink tissues (Fig. 4). CLas cells move to the roots and colonize there to work as source for systemic spread to newly developing tissues like young leaves (Johnson et al. 2014; Pandey et al. 2021; Raiol-Junior et al. 2021). In a greenhouse experiment of 2-year-old Valencia/Cleopatra sweet oranges, several mature leaves that were never exposed to ACPs had positive CLas titers, suggesting that active bacterial movement may occur to allow systemic spread to source tissues (Pandey et al. 2021). In the same study, 65.83% of leaves directly fed on by ACPs, 39.24% of young leaves that emerged post-ACP feeding (minimum age 30 days), and 8.28% of mature leaves not fed on by ACPs displayed blotchy mottled HLB symptom 1 year posttransmission (Pandey et al. 2021). However, the earliest HLB symptom was observed at 3 months post-ACPs feeding (ACPs exposure for 2 to 20 days) in 10-month-old Valencia sweet orange seedlings (Pandey and Wang 2019). CLas inoculation by grafting suggests that the roots of citrus trees are the earliest site of colonization and that the decline in roots occurs prior to the appearance of visible HLB symptom on the shoots and leaves (Johnson et al. 2014). Taken together, the appearance of visible symptoms lags behind CLas transmission, changes in host gene expression, CLas colonization, and CLas titers reaching a threshold that allows them to be taken up by ACPs (Bové 2006; Lee et al. 2015; Pandey et al. 2021; Pandey and Wang 2019; Wang et al. 2017b; Wei et al. 2021). Citrus leaves have an average lifespan of 1.5 years, but some may survive up to 3 years. It takes at most 3 years for CLas to spread and colonize throughout a host and for the host to develop HLB symptoms (Pandey et al. 2021).

Download as PowerPoint

Insights into symptom development.

The symptoms of Ca. Liberibacter spp. caused diseases on citrus and potato show similarity with the symptoms caused by other phloem limiting pathogens such as citrus stubborn disease caused by Spiroplasma citri, potato purple top caused by Ca. Phytoplasma, and potato leafroll virus (Bové 2006; Crosslin and Munyaneza 2009; Lee et al. 2009; Shi et al. 2014; Teixeira et al. 2008; Wang et al. 2017b). These observations suggest that the disruption of phloem functions might be the major contributor of disease symptoms caused by the phloem-infecting pathogens (Wang et al. 2017b).

CLas induced dysfunction of phloem, which is considered one of the drivers of HLB symptom development (Folimonova and Achor 2010; Kim et al. 2009; Koh et al. 2012; Wang and Trivedi 2013). Sieve element blockage, necrosis, and phloem collapse create obstacles to photoassimilates transport, resulting in excessive accumulation of starch in the leaves and stem of HLB-positive plants (Aritua et al. 2013; Fan et al. 2010; Gonzalez et al. 2011; Koh et al. 2012). CLas infection also induces callose accumulation around plasmodesmata pore units connecting the sieve elements with companion cells (Koh et al. 2012). Koh et al. (2012) demonstrated that callose deposition around plasmodesmata pore units begins prior the starch accumulation in the chloroplasts. Callose accumulation results in impaired symplastic movement and delayed transport of photoassimilates (Koh et al. 2012). These results suggest that the obstruction to photoassimilates transport causes the starch accumulation in CLas-positive leaves (Koh et al. 2012). Anatomical aberrations in the phloem like the collapse of sieve elements, companion cells, and occasionally the cambium cells; the swelling of middle lamella; and the plugging of nonnecrotic sieve pores with callose and P-protein–like materials all correlate with the development of HLB symptoms (Folimonova and Achor 2010). Apparently, the ROS generated by CLas colonization triggers the programmed cell death and is primarily responsible for the collapse of sieve elements (Ma et al. 2021). Tolerant citrus varieties show lower disruption of sieve elements and higher phloem regeneration, which may be the key elements to provide tolerance against the HLB (Deng et al. 2019).

The anatomical structure of root tissues from HLB-positive citrus trees were marked by swollen middle lamella, lower starch, enriched cytoplasmic contents, and collapsed sieve element (Aritua et al. 2013). Interestingly, the earliest root decline post-CLas movement and colonization was observed before any foliar symptoms or the reduction in starch content and phloem plugging in roots (Johnson et al. 2014). These results suggest the crucial role of the root system in initial development of HLB symptoms, which may not be associated initially with phloem plugging.

CONCLUDING REMARKS

Since the discovery of HLB pathogens, seven Ca. Liberibacter species have been reported so far. Ca. Liberibacter pathosystems have been shown to affect many important crops. It is anticipated that the advance of new technologies, such as CRISPR and artificial intelligence, presents ample opportunities to further unravel the ambiguity of the Ca. Liberibacter pathosystems, contributing to develop efficient, sustainable, and environmentally friendly approaches to manage these diseases.