Experiments Being Carried Out in Our Laboratory

Sources of virus isolates

Florida CTV isolates T66 sub-isolate E, T36 (quick decline-inducing isolate) and T30 (mild isolate) were maintained at the University of Florida’s Citrus Research and Education Center, Lake Alfred, Florida. Isolate B249 (stem pitting isolate from Venezuela) was obtained from the Collection of Exotic Citrus Pathogens main- tained at the USDA-BARC, Beltsville, Maryland. Isolate DPI 3800 (a mixture of stem pitting-like and T36-like strains from Florida) was acquired from the Florida Department of Plant Industry (DPI), Gainesville, Florida. All isolates were prop- agated on Mexican lime.

Plasmid vector construction and bacterial strains

Since the process of producing and growing transgenic citrus is labour intensive and

rather ineffi cient, and CTV is a complex virus, our strategy has been to introduce individually a wide variety of CTV sequences to increase the chances of pro- ducing tolerant plants. The sequences include the CP gene from three different strains with distinct biological characteris- tics, a non-translatable version of one of the CP genes, the RdRp, p27 and p20 genes, and the 3¢ end (400 3¢ -terminal bases, including part of the p23 gene and the 3¢ UTR) of the CTV genomic RNA. Additionally, we have also thus far used the NPR1 gene from Arabidopsis in transformation experiments in an attempt to obtain broad-spectrum dis- ease resistance. The construction of these vectors has been described in detail else- where (Febres et al., 2003). Four binary vector plasmids were used to clone the sequences for their delivery into grapefruit tissue via Agrobacterium tumefaciens

transformation. These vectors were pGA482GG, pMON10098, pCAMBIA 2201

and pCAMBIA 2202. A summary of the constructs used for the transformation results presented here is shown in Table 14.1.

Pathogen-derived resistance can be limited to the strain from which the trans- gene was derived or to closely related strains (Lomonossoff, 1995). These con- structs are varied in terms of the genes they contain, the position of the genes in the CTV genome (both 3¢ and 5¢ regions are rep- resented) and their origin (virus strain). Since the 3¢ end of CTV is the most con- served region among different strains of the virus (> 97%), we speculated that this region was the most likely to induce resist- ance to the widest varieties of CTV strains. Other sequences (CP and RdRp) were chosen because they have been widely and

Table 14.1. Description of the plasmids and bacterial strains used for the Agrobacterium tumefaciens - mediated transformation of grapefruit.

a35S = CaMV 35S promoter; 34S = FMV 34S promoter; CP T36 = major CP gene from decline-inducing isolate T36; CP T30 = major CP gene from mild isolate T30; CP B249 = major CP gene from stem pitting-inducing isolate B249; RdRp = replicase gene from isolate T36; p20 = repressor of RNA silencing from isolate T36; p27 = minor CP gene from isolate T36; NT CP = non-translatable major CP from isolate T36; 3END-S = 3¢ end (400 3¢ -terminal bases, including part of the p23 gene and the 3¢ UTR) in the sense orientation from stem pitting-inducing isolate DPI 3800; 3END-AS = 3¢ end from isolate DPI 3800 in the antisense orientation; NPR1 = A. thaliana NPR1 gene (non-expressor of pathogenesis- related proteins 1).

bKan = kanamycin.

cGUS = glucuronidase; GFP = green fl uorescent protein.

dTet = tetracycline; Spc = spectinomycin; Cap = chloramphenicol.

successfully used for inducing resistance in other plant–virus systems (Lomonossoff, 1995; Hammond, 1999). The p20 gene was chosen because it is highly expressed and is thought to be important in the virus life cycle, and also it is located in the more con- served 3¢ end of the virus. The sequences for the CP T36, NT CP, RdRp, p27 and p20 were cloned by reverse transcription–poly- merase chain reaction (RT–PCR) from the Florida CTV isolate T36 and were identical to the publicly available sequence for this isolate (gb U16304). Similarly, the sequence for the CP T30 was identical to the public sequence available (gb AF260651). The 3¢ end of CTV was cloned from isolate DPI 3800, a mixture of quick decline and stem pitting strains. The sequences of the 35S- 3END-S and 34S-3END-AS constructs were T36-like, whereas the sequence of the 34S- 3END-S construct was more similar to stem pitting isolates (data not shown). The NPR1 gene from Arabidopsis was cloned into the pCAMBIA transformation vectors from genomic DNA using specific primers designed based on the sequence publicly available (gb U76707.1).

Plant transformation

Grapefruit (C. paradisi Macf. cv Duncan) seeds were extracted from mature fruits and used for A. tumefaciens -mediated transfor- mation according to the procedure previ- ously described (Luth and Moore, 1999). Four- to 6-week-old etiolated stem seg- ments were used as explants. An initial b- glucuronidase (GUS) assay was performed on all of the shoots that regenerated on selection medium (some 5–8 weeks after inoculation with Agrobacterium) by remov- ing small sections from their basal ends, fol- lowed by histochemical GUS staining (Moore et al., 1992). A second GUS assay was performed on leaf segments of rooted plants using a similar staining procedure. Similarly, green fl uorescent protein (GFP) was visualized in intact shoots and plants using a fl uorescence stereoscope fi tted with a 535nm emission fi lter (Seizz). A standard

PCR was performed on the putatively trans- genic plants (positive in the fi rst GUS or GFP assay) using gene-specific primers (Febres et al., 2003).

The plants obtained from the transfor- mation experiments resembled the wild type morphologically and displayed normal growth. The results of the plant transforma- tion experiments with the CTV sequences are summarized in Table 14.2. There was variability in the number of explants (stem segments) regenerating shoots (between 4 and 13%, with an average of 9%), the number of GUS-positive shoots (between 6 and 39%, with an average of 14%) and the number of solid (non-chimeric in the histo- chemical assay) GUS-positive shoots (between 0 and 40%, with an average of 17%). The control plasmid pGA482GG, without any CTV genes, produced an above average number of GUS-positive shoots and the highest number of solid GUS-positive shoots. This could be due to a higher effi - ciency in the transfer of smaller Agrobacterium T-DNA segments or perhaps to a detrimental effect of the CTV genes.

None of the p20-regenerated shoots and only one of the p27-regenerated shoots were PCR positive for the gene of interest (Table 14.3), even though large numbers of stem seg- ments were inoculated with Agrobacterium containing these constructs (Table 14.2). This does not seem to be a problem caused by the ineffi ciency of the pMON10098 vector in comparison with the pGA482GG or pCAMBIA 2201 vectors, since the initial numbers and percentages of GUS-positive shoots produced using the constructs with the p20 and p27 genes was high (19 and 10% of regenerated shoots, respectively, Table 14.2). It is possible that these two genes are toxic or deleterious to the plant and are removed from the genome by rearrange- ments, or that such transgenic plants die at an early stage. This would be consistent with a role for p20 as a repressor of RNA silencing. It is now accepted that RNA silencing is one of the mechanisms used by the cell to regu- late gene expression (Voinnet, 2002). Such a repressor constitutively expressed could alter development in a deleterious way. As for

Table 14.2. Transgenic shoot production from the stem segment explants of grapefruit inoculated with

Agrobacterium tumefaciens carrying different sequences of CTV.

aPercentage refers to the number of segments producing shoots versus the number of segments analysed.

bPercentage refers to the number of GUS+ shoots versus the total number of shoots.

cPercentage refers to the number of solid GUS+ shoots versus the number of GUS+ shoots.

p27, the only known role is that of a struc- tural protein; however, it is not uncommon for viral proteins to have more than one func- tion in the virus life cycle.

After the transgenic plants were estab- lished in soil, we performed a second GUS assay or GFP visualization. Not all of the plants that tested positive in the fi rst assay were positive in the second assay (Table 14.3). In most cases, the second assay GUS/GFP-negative plants originated from the chimeric shoots observed in the fi rst assay (data not shown). We were interested in determining whether the second GUS/GFP assay was a good indication for the presence of the transgenes, compared with the fi rst assay. Excluding the results for 35S-p27 and 35S-p20, most of the GUS/GFP- positive plants in the second assay (> 75%, and in most cases close to 100%, with an average of 91%) were also PCR positive for the gene of interest (Table 14.3). However, a few of the plants (an average of 12%) that were negative in the second GUS/GFP assay (but positive for GUS/GFP in the fi rst assay) contained the gene of interest (Table 14.3). Even though these percentages are low, since the number of putatively transgenic plants

(PCR positive for the gene of interest) is also relatively low, it may be worthwhile to analyse all of the rooted plants by PCR even if they are negative in the second reporter gene assay. Further, these plants may repre- sent a silenced state that could be of more interest for our purpose, at least in the case of the CTV sequences.

A selected number of the putatively transgenic plants were further analysed by Southern blots to corroborate the integration of the genes into the plant genome and to determine their copy number. The results indicated that we obtained stably trans- formed grapefruit plants with one to several copies of the transgene per genome (data not shown and Febres et al., 2003). Further, western blot analysis of the CP transgenic plants showed the presence of the 25 kDa CP in several lines, demonstrating that the CP from some of these lines was expressed to detectable levels (data not shown).

Virus challenge of the transgenic plants

In nature, CTV is transmitted by aphids (Bar-Joseph et al., 1989). Unfortunately, it is

Table 14.3. PCR analysis of the regenerated reporter gene-positive grapefruit shoots using transgene-specifi c primers.

aPercentage refers to the number of PCR-positive shoots versus the number of second assay GUS- positive shoots.

bPercentage refers to the number of PCR-positive shoots versus the number of second assay

GUS- negative shoots.

cResults from GUS- and GFP-positive plants were combined for this analysis.

very diffi cult to infect CTV into a citrus plant by mechanical inoculation (Muller and Garnsey, 1984; Bar-Joseph et al., 1989). Therefore, to test whether the transgenic plants we are producing show resistance to CTV, they are being challenged by both aphid and graft inoculations (Fig. 14.3). The results shown here are mostly from graft inoculation with CTV isolate T66-E. Buds from verifi ed transgenic plants (GUS, PCR and Southern blot positive) were grafted on Swingle citrumelo (Poncirus trifoliata (L.) Raf. ´ C. paradisi Macf.). Each transgenic plant was budded in fi ve replicates, and one of the transgenic plants was maintained for healthy controls. After the transgenic buds began to grow, the plants were graft challenged with CTV severe isolate T66-E by grafting three infected leaf segments on to the transgenic scion. After 12 and 18 months, the transgenic grapefruit scions were tested by double antibody sandwich indirect enzyme-linked immunosorbent assay (DASI-ELISA) (Garnsey and Cambra, 1991) using the polyclonal antibody UF 1052 for coating and the monoclonal anti- body MCA13 as secondary antibody.

MCA13 reacts predominantly with the CP of severe isolates of CTV (such as T66) and does not react with T30 (Permar et al., 1990). Each ELISA sample was tested in duplicate. In the case of the aphid inocula- tions, colonies of Toxopthera citricida (brown citrus aphid) were fed for at least 48 h on T66-E-infected Mexican lime plants. Groups of 20 aphids were transferred to the transgenic plants (additional propagations from those used in the grafting experi- ments) and allowed to feed for several days. Non-transgenic plants were used as con- trols for the effi ciency of virus transmis- sion. Infection by CTV was also tested using ELISA.

Only a small, representative number of all the lines challenged are shown in Table

14.4. Most of the lines did not show resist- ance to CTV (line 81, for example), testing positive for CTV. A few of the lines, how- ever, showed some of the plants free of CTV (lines 146, 169, 212, 538 and 595). Plant 595 (transformed with the 3END construct in the sense orientation) was the most resistant. Interestingly, a couple of the lines showed plants that seemed to have recov-

Fig. 14.3. Grafting procedure used to challenge the transgenic plants. (A) The materials required for graft- ing are a scalpel, blade or sharp knife and grafting tape. If grafting tape is not available, ‘Parafi lm’ can be used instead. However, it tends to ‘melt’ with time and it becomes more diffi cult to remove. (B) The main vein of the leaf is scraped off to expose the phloem tissue. (C) A small square or ‘chip’ of about 3–4 mm wide is cut out from the exposed phloem portion. (D) A cut is made to the stem of the plant to be challenged as shown. Ideally, straight portions of the stem are selected for the graft, away from leaves or other branches, which are removed. (E) The chip is placed between the stem and the fl ap, with the phloem of the chip toward the phloem of the stem. (F) Grafting tape is wrapped around and used to secure the graft in position. In our experiments, we make three to four grafts onto each transgenic scion. After 4–5 weeks, the tape is removed. By then the graft should have taken. The chip should be green and the fl ap eventually dies and falls off.

ered from the viral infection (line 212). Overall, 41 lines of transgenic plants have been tested, of which 13 have shown some degree of resistance. These results are simi- lar to those of Dominguez et al. (2002) in which only a proportion of the plants showed resistance to CTV.

We initiated the effort to produce CTV- resistant citrus several years ago. At the time, little was known about the mecha- nism involved in pathogen-mediated dis- ease resistance. During the time we have been producing and challenging these

plants, a time-consuming and laborious process, much has been learned about what occurs during pathogen-mediated disease resistance and how to bring it about effec- tively (Waterhouse, 2001; Rovere et al., 2002; Voinnet, 2002; Wassenegger et al., 2002a, b; Tang et al., 2003). Any of the genes that show promise in the transgenic plants we have produced will be intro- duced into transgenic plants in constructs that include inverted repeats that result in RNA with a double-stranded ‘hairpin’ structure (Waterhouse et al., 1998), direct

ELISAb

aPlants were challenged with CTV by either grafting (Graft) or aphid transmission (Aphid), or remained unchallenged.

bValues are averages of OD415 from two repetitions estimated 12 or 18 months after inoculation. A positive result is at least twice the average OD of the uninfected controls. Values in bold indicate those

samples considered negative. N/A = sample not assayed.

tandem repeats of three or four copies of the gene (Ma and Mitra, 2002) or short (40–60 bp) inverted repeats (Lecomme et al., 2003). These strategies are likely to be more effi cient (close to 100% silencing) than the single gene strategy followed to this point.

Experiments on natural disease resistance

Of course, CTV is not the only pathogen that plagues citrus; there are a host of others of bacterial, fungal, viroid and viral origin. As described above, mechanisms of plant disease resistance are being elucidated in annual species such as Arabidopsis, and a

number of the genes involved have been isolated in recent years. The most promis- ing characteristic of using resistance genes of plant origin such as NPR1 is that they may lead to broad-spectrum protection against a variety of pathogens. In addition, since they are of plant origin, they should be of less concern to consumers. Our citrus gene isolation and transformation efforts are in their early stages. In the single report where citrus was transformed with a PR protein, a transgenic sweet orange line con- taining a tomato PR-5 gene was signifi - cantly more resistant to the oomycete Phytophthora citrophthora (Fagoaga et al., 2001).

One of our major objectives is to deter- mine whether the genes shown to be impor- tant in disease resistance in other species, such as EDR1, EDS1, EDS5, NDR1, NPR1, PAD4, PBS1, PBS3, PR1, RAR1, SGT1 and

RdRp, are present in citrus. We are using two strategies to do this. One is to BLAST (https://www.ncbi.nlm.nih.gov/BLAST/) the sequence of proteins of known function from other organisms to the Citrus expressed sequence tag (EST) database available to identify potentially homolo- gous genes. In recent years, EST sequences from various citrus species and tissues, and under different biological and environmen- tal conditions have been added to GenBank (https://www.ncbi.nlm.nih.gov), allowing easy access to this information for BLAST comparison and identifi cation of genes of interest. As of August 2007, more than 94, 000 entries were available for C. sinensis (sweet orange) and thousands more for other Citrus types (https://www.ncbi.nlm. nih.gov/dbEST/dbEST_summary.html), and this number is rapidly increasing. Several groups around the world, including Japan, California (UC, Riverside’s HarVest: Citrus, https://harvest.ucr.edu/) and Florida (USDA), among others, are actively adding more EST sequences to the data- base.

Alternatively, degenerate primers based on protein sequence alignments can be designed in those cases where BLAST does not identify any homologous

sequences. A useful tool for the design of degenerate primers is the CODEHOP program (https://blocks.fhcrc.org/codehop.html; Rose et al., 2003).

One of the genes we have cloned using these strategies is the grapefruit EDR1 homologue. This Citrus gene has the kinase domain and the conserved regions present in EDR1 orthologues but is diver- gent in paralogues (Fig. 14.4). As men- tioned above, this gene is a negative regulator of SAR. A mutated, kinase-defi - cient version was transformed into Arabidopsis, and the regenerated plants showed increased resistance to powdery mildews (Tang and Innes, 2002). We antici- pate that a similar strategy could work in citrus.

Because citrus is affected by such a diversity of pathogens, fi nding individual genetic cures for each one of them is time consuming and ineffi cient. Ideally broad- spectrum resistance is desired, and the use of the same natural defence mechanisms that are present in plants has the potential of being very effective, based on the limited available evidence. The idea behind using genetic engineering is to produce citrus plants that have a heightened or much ear- lier defence response than wild types after recognition of a pathogen, stopping it short of infection. Most of the experiments on genetically engineered resistance against pathogens have been developed in annual crops. In such cases, just a delay in the onset of a disease can be enough to guaran- tee a profi table production. In the case of a perennial crop, such as citrus, this is not enough. In our case, we want to produce plants that are either immune or highly tol- erant to pathogen infection over long peri- ods of time.

In summary, some of the defence genes that have been identifi ed in other species and that can provide broad-spectrum resist- ance appear to be able to act in heterologous systems. However, we are in the process of identifying some of these genes from Citrus types as well. We intend to test both het- erologous and homologous genes in trans- genic plants to determine their potential in

Fig. 14.4. Amino acid sequence alignment of the kinase domain of EDR1-type proteins from several plant species including grapefruit (Cp). EDR1 is a Raf-like mitogen-activated protein kinase kinase (MAPKK) that functions as a negative regulator of disease resistance and SAR. Horizontal bars indicate regions conserved in EDR1 orthologues but divergent in paralogues (CTR1). Cp = Citrus paradisi; Le = Lycopersicon esculentum; At = Arabidopsis thaliana; Hv = Hordeum vulgare; Os = Oryza sativa.

inducing long-lasting, broad-spectrum dis- ease resistance against the pathogens that plague this important crop.