Source, identity, and naming of plant material.

Seeds for all progenies were provided by the USDA–Agricultural Research Service (ARS) National Clonal Germplasm Repository for Citrus and Dates at UCR, and were harvested from fruit of field-grown trees in the CVC of UCR (see http://www.citrusvariety.ucr.edu). Accessions in the CVC are designated with CRC numbers. The seedling progenies of 84 seed-source accessions were used in this study and are referred to by the CRC number of the seed parent accession. The online version of Table 1 provides access through hyperlinks to the history of each accession with taxonomic references, from the CVC (CRC) and the Germplasm Resource Information Network–Global of the USDA-ARS National Plant Germplasm System. Within the genus Citrus, incidence of nucellar embryony varies greatly [reviewed in Frost and Soost (1968)]; therefore, some of the plants we tested were essentially identical genetically to the female parent (seeds serving as clonal propagules), whereas others were half-sib hybrids of known seed parents but with undetermined pollen parents. As described later, the vast majority of these trees are expected to result from nucellar embryony. For convenience, they are referred to as “accessions,” using the CRC numbers of the seed source accessions. Hereafter, the seed source accessions providing the progeny used in this study are referred to as “parent genotypes.”

Table 1.

Incidence and severity of asian citrus canker (ACC) in leaves of half progenies of 84 seed source accessions exposed to field infection in Ft. Pierce, FL, for 3 years of assessment. In each column, to assist visualization, means with solid gray fill are in the least ACC-sensitive statistical category and are significantly different from those in bold.

View Table

View Table

Plant development and maintenance.

Seedling development, field planting, and general care were conducted as described in Richardson et al. (2011) and Westbrook et al. (2011), and were typical for citrus production, except with minimal pesticide application. Before field planting, seedlings were maintained in a greenhouse free of asian citrus psyllid [vector of Candidatus Liberibacter asiaticus (CLas), causal agent of HGB] and other arthropod pests. Plants were sown for germination in Jan. 2012, with some resown in June 2012 where germination was initially low. As a component of the HLB-related investigation, half of the seedlings from each parent genotype were bud-inoculated with CLas in the greenhouse in Oct. 2012, and all were field-planted during Aug. 2013. In our study, plants were evaluated for ACC 5 to 8 years after germination.

Trees were typically sprayed every 2 weeks with 435 horticultural oil and 200 ppm Magna-Bon (Magna-Bon II LLC, Okeechobee, FL). In a few instances, spraying was suspended for 2 months or more for insect observations in other studies. The nutritional component varied throughout the year and included a wide range of products. The list of materials included 20N–4.4P–16.6K, 14N–0.9P–11.6K, Keyplex 1400 DP (Keyplex, Winter Park, FL), Ocean Organics seaweed extract (Ocean Organics, Ann Arbor, MI), and N-Sure 28N–0P–0K (Tessenderlo Kerley, Inc., Phoenix, AZ). Other nutrients were included periodically, but none that were considered to have any effects on canker.

Assessing ACC sensitivity.

Formation of foliar lesions indicative of ACC is affected by stage of leaf development when exposed to Xcc (Gottwald and Graham, 1992; Graham and Gottwald, 1991; Vernière et al., 2003). Wind-driven rain is the main factor in field spreading of inoculum (Graham et al., 1992), and the combination of conducive environmental conditions and available inoculum must coincide with foliage susceptibility for ACC development. For these reasons, we assessed ACC presence as both incidence (percentage of leaves displaying lesions) and severity (the percentage of leaf area affected by lesions). ACC symptoms were assessed in Aug. 2017, Sept. 2019, and Sept. 2020. The visual estimates of leaf area disease (ACC severity) were guided by a standard area diagram set prepared using image analysis as described in another study (Bock et al., 2010). In 2017, the overall percentage of leaf area displaying ACC lesions on each tree was assessed but was found to display little mean separation; so, in 2019 and 2020, the percentage of leaf area displaying ACC lesions only on the most severely affected leaf was assessed on each tree.

Statistical analyses.

We tested whether the incidence and severity of ACC differed among progenies of the 86 parent genotypes using Kruskal Wallis nonparametric analysis (PROC GLIMMIX, SAS 9.4; SAS Institute Inc., Cary, NC). Regressions were conducted using Microsoft Excel (Microsoft Office version 2016; Microsoft Corporation, Redmond, WA).

qPCR verification of pathogen associated with symptoms.

Verification that purported ACC lesions were in fact a result of infection with Xcc was conducted by qPCR analysis. Four trees were selected randomly for sampling of well-known accessions from each of pure C. trifoliata, citranges, citrumelos, US-119 (the only unifoliate hybrid in the trial), and four unifoliates from adjoining rows. Four leaves showing clear canker lesion symptoms were sampled from different branches on each tree. Lesions and the surrounding area, ≈200 mg of tissue, were excised and macerated for DNA extraction. Carrizo and Hamlin leaves from clean greenhouse trees, extracted in the same manner, were used as negative controls.

DNA was isolated using the ZR-Duet Quick-DNA/RNA Miniprep Plus Kit (Zymo Research, Irvine, CA) according to the manufacturer’s instructions. DNA concentration and purity were analyzed using a NanoDrop 8000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA). qPCR was performed using the PowerUp SYBR Green master mix (Thermo Fisher Scientific) and 100 ng of template on the Applied Biosystems AB7500 Real-Time qPCR system (Thermo Fisher Scientific) according to the manufacturer’s instructions. Primer sets (Cubero and Graham, 2002) were specific to the 16s ribosomal DNA sequence of X. euvesicatoria pv. citrumelonis [abbreviated hereafter as Xec; recent synonyms X. alfalfae pv. citrumelonis (Constantin et al., 2016), X. axonopodis ssp. citrumelonis, and X. campestris pv. citrumelo (National Center for Biotechnology Information, 2020)], lrp gene conserved in Xanthomonas spp., pth gene conserved in Xcc strains, and citrus Glyceraldehyde 3-phosphate dehydrogenase as a reference gene. DNA was isolated from liquid cultures of Xcc (strain ID 3213 from D. Gabriel) and Xec (strain ID 3048 from D. Gabriel) using phenol, phenol–chloroform–isoamyl alcohol (25:24:1), followed by two 70% ethanol washes. The dried pellet was then resuspended in 100 µL 10 mm Tris buffer (pH 8.5) for use as positive controls. In addition to the clean citrus DNA controls described, standard water controls were also included.

Single nucleotide polymorphism analysis of chloroplasts.

To determine chloroplast identities, comparisons were made using data on 462 well-distributed chloroplast single nucleotide polymorphisms (SNPs) included on a 50,000 + SNP chip developed at UCR with polymorphic SNPs selected based on sequences of 41 diverse accessions from the UCR CVC and public databases (Hiraoka, 2020).

ACC pathogen widespread in trial.

ACC was ubiquitous in the research farm where this experiment was conducted. The lesions caused by Xcc infection initially have a distinct yellow halo, which often becomes necrotic, and the lesions feel sandpapery on the abaxial and adaxial surfaces of the leaves, and raised lesions form on the stems as well (Das, 2003). These traits were displayed in all trees throughout the planting. Virtually all lesions had progressed beyond the yellow halo stage in the trial planting, even though neighboring unifoliate citrus—citrus with a leaf composed of a single leaflet rather than the compound trifoliate (3-leaflet) leaf of C. trifoliata and its F1 hybrids—maintained many halos. The bacterium Xec, causal agent of Citrus bacterial spot, is also present in Florida and produces similar symptoms, but is generally reportedly prevalent only in nurseries and is associated with frequent mechanical injury (budding and pruning) to trees [reviewed in Graham and Gottwald (1991]. More recently, there are unpublished reports that the HLB-tolerant cultivar LB8-9 (SugarBelle©) is highly susceptible to Xec but not Xcc, but lesions are only found on leaves and never fruit (F.G. Gmitter, Jr., personal communication). For our study, qPCR with primers to distinguish the two Xanthomonas species (data not shown) was run on ACC-symptomatic leaves from the test plot and adjoining rows, and all tested strongly positive for Xcc [all but two with a cycle threshold (Ct) <25, one at Ct 30.9 and one at Ct 26.0 (mean, 22.6), reflecting a mean copy number of 240,000 per ng plant DNA]; only 44% had detectable Xec and then at very low titers (Ct, >35.4; a mean copy number of 4 per ng DNA). Therefore, the distinctive canker lesions were attributed to Xcc.

ACC sensitivity in accessions.

CLas inoculation before planting had no effect on ACC, thus all plants of an accession were assessed as a single group (data not shown). Almost all accessions with a lower ACC lesion incidence were C. trifoliata hybrids rather than pure C. trifoliata in all 3 years of data collection (Table 1). To aid in visualizing the responses across all accessions, pure C. trifoliata accessions are in bold in the leftmost column of Table 1; in each year for incidence and severity of ACC, the accession means that separated statistically are shaded in gray for low incidence and in bold for high incidence.

Incidence of ACC across parent genotypes ranged from 8% to 80% (mean, 49%) of leaves affected in 2017, from 4% to 58% (mean, 29%) in 2019, and 8% to 46% (mean, 25%) in 2020. In 2017, of 49 C. trifoliata parent genotypes, only four separated from the two highest ACC-incidence statistical categories (CRC 3345, 3484, 3888, and 4017), whereas 29 of the 35 C. trifoliata hybrids displayed lower ACC incidence, which separated from the two highest statistical categories (the two highest were used for this comparison because so few accessions were in the highest group). In 2019, of the C. trifoliata, only six separated from the highest ACC-incidence statistical category (CRC 3330, 3484, 3547, 3549, 3876, and 3888), whereas all 35 C. trifoliata hybrids displayed lower ACC incidence and separated from the highest statistical category, and 26 hybrids separated from 18 of the C. trifoliata. In 2020, only five C. trifoliata separated from the highest ACC-incidence statistical category (CRC 2861, 3484, 3549, 3888, and 4008) and 20 hybrids separated from 18 of the C. trifoliata. C. trifoliata accessions CRC 3484, 3549, and 3888 are noteworthy for their separation from the highest ACC-incidence category in 2 or 3 years. Only CRC 3888 had this distinction in all 3 years, but it should be noted that there was only one tree surviving and, because its parent accession is reported to be monoembryonic, it is likely a self-pollinated seedling or a hybrid. It was among the lowest accessions in canker sensitivity by all metrics in all 3 years and merits further investigation. We will propagate this one tree and test for ACC sensitivity in a replicated trial against appropriate standards. It is reported that in open-pollinated seedlings of C. trifoliata accessions at the UCR CVC, all zygotic seedlings were selfs or crosses between C. trifoliata accessions, likely a result of earlier flowering compared with other citrus (Kepiro and Roose, 2009). If this is a self or a hybrid with another trifoliate, it could prove useful in understanding more fully the genetics of ACC sensitivity.

By accession means, ACC incidence correlated between each pair of the 3 years, with r² values of 0.39, 0.57, and 0.65 (Table 2). Within each year, metrics of ACC severity correlated with incidence of ACC, with r² values of 0.26, 0.94, and 0.42 for 2017, 2019, and 2020, respectively.

Table 2.

r² Values from regressions of incidence and severity of asian citrus canker (ACC) in leaves of half progenies of 84 seed source accessions exposed to field infection in Ft. Pierce, FL, for 3 years (2017, 2019, and 2020) of assessment. Statistical significance of all regressions exceeds P < 10^–7.

View Table

ACC sensitivity within citrumelos and citranges.

Citrumelo, the name given to hybrids of grapefruit and trifoliate orange, is widely used as citrus rootstock, and ‘Swingle’ citrumelo has been the most widely used citrus rootstock in Florida for the past 30 years (Florida Department of Agriculture and Consumer Services, 2016, 2020). Three of the 10 citrumelo accessions in our trial are independent entries of ‘Swingle’ (CRC 3341, 3771, and 3767), which were verified by SNP analysis, and all were consistently among the least ACC-sensitive accessions (Table 1). Most citrumelo accessions did not separate from the least ACC-sensitive category across all 3 years (9 in 2017, 8 in 2019, and all 10 in 2020). Two citrumelos were among the most ACC-sensitive accessions in this trial in some but not all years (CRC 3348 in 2017, and CRC 3821 in 2019 only), showing significant mean separations for ACC incidence. All citrumelos in the trials had a grapefruit seed parent.

Citrange, the name given to hybrids of sweet orange and trifoliate orange, is also widely used as citrus rootstock, and ‘Carrizo’/‘Troyer’ (shown to be the same genotype) is the most widely used citrus rootstock in California and second-most widely used in Florida for the past 30 years (Florida Department of Agriculture and Consumer Services, 2016, 2020). Of the 14 citrange accessions, ‘Morton’ and ‘Rusk’ were duplicated, and ‘Carrizo’ matched ‘Troyer’ by SNP analysis (Roose unpublished). By chloroplast analysis, 10 accessions had a sweet orange seed parent and four trifoliate orange. In no year and by no metric were any citranges statistically different from each other in ACC sensitivity (Table 1). Only in 2017 did any citrange fail to fall in the least ACC-sensitive statistical grouping (CRC 271 and 275).

ACC sensitivity by C. trifoliata subgroup.

There are four near-isogenic groupings of C. trifoliata accessions, designated as restriction fragment length polymorphism (RFLP) types 1 through 4 by Fang et al. (1997), based on an array of molecular methods, and more recently verified by high-density SNP analysis (Roose, unpublished). Among the less ACC-sensitive C. trifoliata accessions, CRC 3888 (lowest sensitivity in all 3 years) and CRC 4017 are the only two accessions in this study from the uncommon RFLP type 2, whereas CRC 3484 and CRC 3549 are in the major group RFLP type 3, corresponding to the established “small flower trifoliate” group. For the other common C. trifoliata designated as RFLP type 4 (corresponding to large flower trifoliates), only CRC 3345 displayed ACC incidence that separated from the two most severe statistical groupings, and only in 2017. CRC 3938 was the only RFLP type 1 accession, and it was in the highest ACC sensitivity group in all 3 years for all metrics. Statistical analysis of all C. trifoliata accessions by RFLP types showed type 2 to be significantly less sensitive than other types for ACC incidence only in 2017 (Table 3). A separate analysis comparing types 3 and 4 only (because they were much more numerous) showed type 3 accessions to be less sensitive by metrics of ACC severity in all 3 years and by ACC incidence only in 2019.

ACC severity showed less robust mean separation in all 3 years (Table 1) and seemingly provided no noteworthy additional inferences. Because only a few accessions displayed statistically different means in 2017, these data are excluded from Table 1, but are presented here: Citrumelo CRC 1452 and trifoliates CRC 1717, 3338, 3548, 3570, 3586, 4006, and 4008 had a significantly greater percentage of canopy covered by lesions (10.6% to 13.9%) than trifoliate CRC 3587 (5.0%).

ACC sensitivity as affected by chloroplast type.

The UCR database on SNPs permitted assessment of the actual seed parent of the hybrids, as the cytoplasm is largely derived from the maternal parent in citrus (Abkenar et al., 2004). Eleven of the hybrids were the reciprocal cross from that expected, with the “female first” listing rule not observed. These are indicated by bold text in the “Seed parent (by chloroplast genome)” column (Table 1). ‘Yuma’ was previously found to be a citrumelo rather than a citrange (Xiang and Roose, 1988), and has a grapefruit chloroplast. Of the 34 hybrids with chloroplasts validated, similar numbers of field seedlings had trifoliate, grapefruit, and sweet orange chloroplasts (74, 91, and 91, respectively), permitting the assessment of chloroplast influence on ACC sensitivity. The analyses were conducted in two ways: 1) by evaluating trifoliate vs. grapefruit vs. sweet orange chloroplasts separately or 2) by combining sweet orange and grapefruit chloroplasts into a “pummelo” chloroplast category, because both sweet orange and grapefruit chloroplasts are known to be of pummelo origin (Nicolosi et al., 2000), although they are somewhat different from each other (Carbonell-Caballero et al., 2015). In all years, chloroplast type significantly but modestly affected ACC incidence and/or severity, but not in a consistent manner (Table 4). This analysis was conducted because Omar et al. (2017) reported that replacing the grapefruit chloroplast with that of ‘Meiwa’ kumquat reduced ACC sensitivity markedly. ‘Meiwa’ is markedly ACC insensitive, but sweet orange is only somewhat less ACC sensitive compared with grapefruit and trifoliate orange.

Table 3.

Association with chloroplast type for incidence and severity of asian citrus canker (ACC) in leaves of half-sib progenies of 84 seed source accessions exposed to field infection in Ft. Pierce, FL, for 3 years (2017, 2019, and 2020) of assessment.

View Table

Table 4.

Association with restriction fragment length polymorphism (RFLP) type for incidence and severity of asian citrus canker (ACC) in leaves of progenies of 49 seed source C. trifoliata accessions exposed to field infection in Ft. Pierce, FL, for 3 years (2017, 2019, and 2020) of assessment.

View Table

ACC sensitivity comparison with previous studies.

In an earlier study of more diverse citrus accessions, we reported incidence of ACC in Sept. 2010 and 2012 (Stover et al., 2013). That study included two shared C. trifoliata accessions (CRC 4007 and 3549) and three shared C. trifoliata hybrid accessions (CRC 301, 3771, 3957), but results were rather different. The only statistically significant difference among these accessions in 2010 and 2012 was that CRC 3771 (‘Swingle’ citrumelo) had a significantly greater reported ACC incidence (Sept. 2010, 62% of leaves vs. 34% to 37%; Sept. 2012, 28% vs. 6% to 11%) than the other accessions, and indeed was among the highest in the entire trial. In contrast, in 2017 and 2019, CRC 3771 was significantly lower in canker incidence compared with one or both of the shared C. trifoliata accessions, and lower in severity compared with both (2019 only) of these C. trifoliata accessions. As noted earlier, there are two strains of Xanthomonas reported to be in Florida that infect citrus and produce similar symptoms in C. trifoliata and its hybrids (Graham et al., 1990). Infections by the strain of X. campestris pv. citri (now Xcc) were reported to be less severe in C. trifoliata than several other citrus types, including ‘Swingle’ citrumelo, whereas an aggressive strain of X. campestris pv. citrumelo (now Xec) was most aggressive in ‘Swingle’, followed closely by C. trifoliata, and was much less severe in other citrus genotypes tested. In contrast, one of the Xec strains was significantly more severe in C. trifoliata than in ‘Swingle’ citrumelo, which in turn showed greater lesion diameter than six of seven other citrus genotypes, and only Duncan grapefruit was similar to C. trifoliata in response. In another Florida field screening of ACC susceptibility, ‘Swingle’ citrumelo was reported to be the most susceptible of 10 rootstock and 8 scion genotypes, and this was adjacent to highly infected grapefruit (Futch and Graham, 2012). It seems reasonable that the greater susceptibility of C. trifoliata compared to its hybrids, in the trials reported here, is a result of strain differences, perhaps influenced by the planting being all C. trifoliata and hybrids rather than being predominately unifoliate citrus. Populations of the most aggressive strain of Xec were reported to be sustained after initial inoculation in C. trifoliata and hybrids, but not in other citrus types tested, in which the inoculum was undetectable at 40 d (Graham et al., 1990). Because spread and establishment of the canker pathogen is largely dependent on wind-driven rain and the stage of leaf development when exposed, weather conditions when leaves are susceptible can have a major impact on disease development (i.e., Bock et al., 2014)

Literature support for predominance of nucellar embryony.

In unpublished work, 50 seeds of each of 68 C. trifoliata accessions were evaluated visually for polyembrony (D. Hutchinson, personal communication), and 61 of them displayed greater than 80% of seeds as polyembryonic. Of these accessions, 24 were also included in our study and all were highly polyembryonic. Similarly, in another study, 14 C. trifoliata accessions were included in a larger test of citrus accessions, and 12 were predominately polyembryonic (Krueger and Roose, 2003). However, somewhat greater incidences (≈30%) of zygotic embryo production are reported for some C. trifoliata hybrids with sweet orange and grapefruit parents (Xiang and Roose, 1988), and in crosses between 100% monoembryonic C. maxima × C. trifoliata, 92% of the progeny population produced fewer than 30% polyembryonic seeds (Kepiro and Roose, 2009).