Toxicity to Eldana saccharina of a recombinant Gluconacetobacter diazotrophicus strain carrying a truncated Bacillus thuringiensis cry 1 Ac gene

The stalk borer Eldana saccharina Walker (Lepidoptera: Pyralidae) is a serious pest of sugarcane in Africa. Because E. saccharina larvae feed inside the stem of sugarcane, expression of insecticidal Bacillus thuringiensis Cry proteins in sugarcane endophytes, such as Gluconacetobacter diazotrophicus, would enable the Cry proteins to be produced in the feeding zones of burrowing larvae. To evaluate the potential of using a Cry-expressing G. diazotrophicus strain for the control of E. saccharina, a truncated B. thuringiensis cry1Ac gene was cloned into a G. diazotrophicus strain and the toxicity of the recombinant strain to E. saccharina larvae was evaluated. Bioassays showed that the recombinant G. diazotrophicus strain carrying the cloned cry1Ac gene caused significantly higher E. saccharina mortality and larval growth inhibition than the non-recombinant G. diazotrophicus strain. This study sets a foundation for E. saccharina control strategies that are based on the expression of truncated cry1Ac genes in G. diazotrophicus.


INTRODUCTION
Bacillus thuringiensis (Bt) is an aerobic, spore-forming, Gram-positive bacterium that synthesizes one or more parasporal proteinaceous crystals during sporulation (Höfte and Whiteley, 1989;Schnepf et al., 1998).The Bt crystals may consist of either Cyt proteins or Cry proteins (Höfte and Whiteley, 1989;Schnepf et al., 1998).As a group, Cry proteins are toxic to certain species in the orders Lepidoptera, Diptera, Coleoptera, Hymenoptera, and Hemiptera, and also to other invertebrates such as nematodes and mites (Schnepf et al., 1998;van Frankenhuyzen, 2009).Due to desirable characteristics, such as low toxicity to non-target organisms (Höfte and Whiteley, 1989), Bt insecticidal crystals have been used in several commercial products that are formulated for spray application using conventional spray equipment (Luttrell et al., 1999;Schnepf et al., 1998).Since Bt bioinsecticides do not have systemic activity in plants, they are not effective against insects that burrow into the host plant and feed in the stem.
The stalk borer Eldana saccharina Walker (Lepidoptera: Pyralidae) is the most destructive pest of sugarcane in southern Africa and causes severe sugarcane crop losses (Way, 2004).E. saccharina bores into the sugarcane stem and feeds internally for most of its life cycle.Primarily due to this feeding behaviour, Bt crystal-based formulations have not been effective against E. saccharina in sugarcane fields (D. Conlong, South African Sugarcane Research Institute, personal communication).
In an attempt to improve the efficacy of Bt against lepidopteran larvae that burrow into the sugarcane stem, considerable research effort has gone into alternative delivery strategies, such as the development of transgenic sugarcane or the use of plant-associated bacteria (PAB) as delivery vehicles for Cry proteins.Although transgenic sugarcane plants expressing Cry proteins have been constructed and tested against some sugarcane pests (Arencibia et al., 1997), a review of the literature revealed no published papers that evaluated Cry-expressing sugarcane for the control of E. saccharina.Both epiphytes (bacteria that live on the surface of plants) and endophytes (bacteria that inhabit, for at least one period of their life cycle, the interior of a plant) have been used as delivery vehicles for Cry proteins (Bora et al., 1994;Lampel et al., 1994;Salles et al., 2000).To target insects that burrow into a plant, endophytic bacteria are likely to be better vehicles for Cry proteins than epiphytic bacteria.
Gluconacetobacter diazotrophicus (Acetobacteraceae) is an obligate aerobic, Gram-negative, endophyte that has been isolated from the roots and stems of sugarcane (Gillis et al., 1989;Reis et al., 1994).Primarily due to its ability to fix atmospheric nitrogen, G. diazotrophicus has been the focus of intense study (Dong et al., 1997;Gillis et al., 1989;James et al., 1994;Lee et al., 2000;Reis et al., 1994;Saravanan et al., 2008;Sevilla and Kennedy, 2000).Salles et al. (2000) evaluated the potential of using G. diazotrophicus strain BR11281 as a vector to express a Bt cry3A gene for the control of coleopteran pests of sugarcane.Although the authors did not evaluate the insecticidal activity of the recombinant G. diazotrophicus strain, the study of Salles et al. (2000) suggested that G. diazotrophicus may have potential as a vector to express Bt cry genes for the control of sugarcane pests.
To our knowledge, no published studies have evaluated G. diazotrophicus as a vector to express lepidopteran-specific cry genes for the control of lepidopteran insects, such as E. saccharina.Because G. diazotrophicus inhabits the sugar solution in the intercellular-space apoplast of the stem cortex of sugarcane (Dong et al., 1997), it is a good candidate for the creation of a recombinant endophyte for the control of E. saccharina larvae.In this study, a novel Bt cry1Ac gene construct was cloned into a G. diazotrophicus strain and the toxicity of the recombinant strain to E. saccharina larvae was evaluated.

Bacterial strains and culture conditions
Bt isolate SB-4 (SB-4; cry1Ac gene GenBank accession number AY225453), which was previously isolated from E. saccharina larval cadavers and showed high insecticidal activity against E. saccharina larvae (G.Bouwer, unpublished study), was cultured at 30°C in Luria-Bertani (LB) medium.E. coli strains were cultured at 37°C in LB medium.G. diazotrophicus PAl 5 (hereafter referred to as PAl5) (ATCC 49037 T ) was obtained from the American Type Culture Collection (ATCC) (Manassas, USA) and was cultured at 30°C in the medium described by James et al. (1994), hereafter referred to as JM medium.

General molecular techniques
General molecular techniques, such as ethanol precipitation of DNA and agarose gel electrophoresis (AGE), were performed as described by Sambrook et al. (1989).DNA was quantified with a DyNA Quant 200 Fluorometer (Amersham Biosciences, Piscataway, NJ) using Hoechst 33258 dye (Sigma, St. Louis, USA) and Calf Thymus DNA (Amersham Biosciences) as the standard.The DNA molecular weight marker used throughout the study was the 1 Kb plus DNA Ladder (Invitrogen, Carlsbad, USA).Plasmids were isolated from E. coli strains by using either a Plasmid Midi Kit (Qiagen) when large yields were required or by using a QIAprep Spin Miniprep Kit (Qiagen) for routine analysis or when small yields would suffice.Restriction enzymes were supplied by Roche Applied Biosciences (Indianapolis, USA).PCR products or linearized plasmids were prepared for cloning by electrophoretic separation in 0.7% (w/v) low melting temperature agarose gels (SeaPlaque GTG Agarose, Lonza Rockland, Rockland, USA), excision of the band of the desired size, and extraction of the DNA from the gel using a QIAquick Gel Extraction Kit (Qiagen, Valencia, USA).Commercial kits and equipment were used as per manufacturer instructions.

Cloning system
The Gateway Cloning System (Invitrogen) was used.The Gateway Cloning System is based on the site-specific recombination reactions of bacteriophage λ in E. coli.All Gateway reaction mixtures (BP or LR) were transformed into Library Efficiency DH5α competent cells (Invitrogen) and, after phenotypic expression in S.O.C. medium, transformed cells were identified on the basis of colony formation on LB plates containing the appropriate antibiotic.The plasmids from transformants were purified (as described above) and analysed by digestion with EcoRI and AGE.

PCR amplification of cry1Ac for cloning
An insecticidal crystal gene of SB-4 was previously characterized (G.Bouwer, unpublished study) and classified as a cry1Ac gene (GenBank accession number AY225453).Using conserved regions in the SB-4 cry1Ac gene, a primer set was designed to amplify the first (starting with "A" in the ATG start codon) 1908 bp of the cry1Ac gene.The forward primer (cry1Ac-attB1) had the following sequence: 5'-GGGGACAAGTTTGTACAAAAAAGCAGGCTTCGAAGGAGATAG AACCATGGATAACAATCCGAACATCAATG-3'.
In addition to a sequence complementary to the 5' end of the cry1Ac gene, the forward primer also has a sequence that contains a ribosome recognition sequence (RRS) and an attB1 sequence which is required for cloning using the Gateway vectors.The reverse primer (cry1Ac-t-attB2) had the following sequence: 5'-GGGGACCACTTTGTACAAGAAAGCTGGGTCCTACCCTAG TTGGTTTGTAGACG-3'.
In addition to a sequence that is complementary to the cry1Ac gene sequence (from position 1889 to 1908), this primer also contains a sequence that codes for a stop codon and an attB2 sequence.DNA analysis and primer design were done using DNAMAN version 4 (Lynnon BioSoft, Pointe-Claire, Canada).
The Expand High Fidelity PCR System (Roche Applied Science), a system that incorporates a proofreading polymerase, was used for PCR.PCR was performed on a PCR-Express Thermocycler (Thermo Hybaid/Thermo Scientific, Rockford, USA).The program for the primer set was 2 min at 94°C followed by 30 cycles of 15 s at 94°C, 30 s at 55°C and 2 min at 72°C, and a final single extension cycle of 7 min at 72°C.Twenty nanogram of SB-4 plasmid DNA, purified using a Plasmid Midi Kit, was used as the template and all other reaction concentrations and volumes were as per the Expand High Fidelity PCR System manual.Amplification was confirmed by AGE, with PCR product sizes compared to the molecular weight markers.

Cloning of truncated cry1Ac into an expression vector
The basic outline of the protocol used to clone the truncated cry1Ac gene construct (including the RRS and stop codon; the construct is hereafter referred to as cry1Ac-t) into an expression vector is provided below.The PCR product (cry1Ac-t-attB) obtained with the primer set was cloned into pDONR201 (Invitrogen) using the PCR Cloning System and the Gateway BP Clonase enzyme mix.The BP reaction mixture was transformed into DH5α and transformants were selected on LB plates containing 50 µg/ml kanamycin (Km).The resultant construct was pENTR-cry1Ac-t.Cloning of cry1Ac-t-attB into pDONR201 was performed to flank cry1Ac-t with attL sites, thus enabling cry1Ac-t to be cloned into Gateway attR-containing expression vectors using the LR reaction.The cry1Ac-t construct was cloned from pENTR-cry1Ac-t into pDEST14 (T7 promoter expression vector; Invitrogen) using the Gateway E. coli Expression System and the Gateway LR Clonase enzyme mix.The LR reaction mixture was transformed into DH5α and transformants were selected on LB plates containing 100 µg/ml ampicillin (Ap).The resultant construct was pEXP14-cry1Ac-t.The cloning of cry1Ac-t into pDONR201 (creating pENTR-cry1Ac-t) and subsequent transfer into pDEST14 (creating pEXP14-cry1Ac-t) was confirmed by EcoRI digestion and analysis of the fragments by AGE.
pEXP14-cry1Ac-t, with cry1Ac-t located downstream of a T7 promoter, was transformed into E. coli BL21-SI competent cells (Invitrogen) and, after phenotypic expression in S.O.C. medium, transformed cells were identified on the basis of colony formation on LBON (LB without salt; Invitrogen) plates containing 100 µg/ml Ap.The BL21-SI strain carrying pEXP14-cry1Ac-t is hereafter referred to as BL21-SI(pEXP14-cry1Ac-t).

Analysis of protein expression in E. coli
BL21-SI(pEXP14-cry1Ac-t) cells were used for the E. coli protein expression studies.The protein expression studies were performed as per the protocol provided with the BL21-SI competent cells (Invitrogen), with LBON (containing 100 µg/ml Ap) as the culture medium.Proteins in cell pellets were analysed by sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE) according to the method of Laemmli (1970), as modified by Davis et al. (1986).The molecular weights of the proteins were estimated by comparing their relative mobilities to those of calibration proteins (Roche Applied Science).Gels were fixed and stained with Coomassie Brilliant Blue R250 (Sigma).

Conversion of plasmid pKT230 into a Gateway destination vector
Plasmid pKT230 (Bagdasarian et al., 1981) was obtained from the Netherlands Culture Collection of Bacteria (NCCB) (Utrecht, The Netherlands).The rest of the vector conversion was performed using Gateway Cloning Technology and the Gateway Vector Conversion System (Invitrogen).The basic outline of the procedure is provided below.A Gateway attR cassette (a cassette with the attR sites used in Gateway destination vectors), was ligated with the linearized (SmaI digestion), dephosphorylated pKT230 using T4 DNA Ligase (Promega, Madison, USA) in ligase buffer.The ligation Rapulana and Bouwer 1209 reaction was transformed into Library Efficiency DB3.1 competent cells (Invitrogen) and, after phenotypic expression in S.O.C. medium, transformed cells were identified on the basis of colony formation on LB plates containing 30 µg/ml chloramphenicol (Cm).
The plasmids from transformed cells were analysed with EcoRI digestion.The resultant construct was named pDESTN230.

Cloning of truncated cry1Ac into pDESTN230
In accordance with Invitrogen's recommendations, pDESTN230 was linearized (NotI digestion) in preparation for cloning.The truncated cry1Ac gene was cloned from pENTR-cry1Ac-t into pDESTN230 using the Gateway LR Clonase enzyme mix.During the LR reaction, cry1Ac-t was transferred via site-specific recombination from pENTR-cry1Ac-t to pDESTN230 (creating pEXP230-cry1Ac-t).The LR reaction mixture was transformed into DH5α and transformants were selected on LB plates containing 40 µg/ml streptomycin (Sm).pEXP230-cry1Ac-t, with the truncated cry1Ac gene under control of the pKT230 Km r promoter, was used to transform PAl5.Because there is a single EcoRI site in pDESTN230 and two EcoRI sites in cry1Ac-t, the cloning of cry1Act was confirmed by the presence of three fragments (approximately 10.4, 2.8, and 0.7 kb) after EcoRI digestion.

Transformation of G. diazotrophicus PAl5 with pEXP230-cry1Ac-t
PAl5 was cultured in JM medium and made electrocompetent using the Procedure for High Efficiency Electro-transformation of E. coli -Preparation of Cells protocol provided in the Pulse Controller Operating Instructions and Applications Guide (Bio-Rad, Hercules, USA).Electroporation was performed with a Gene Pulser fitted with the Pulse Controller accessory (Bio-Rad).The Gene Pulser was set to 25 µF and 2.50 kV, and the Pulse Controller was set to 400 Ω.PAl5 (200 µl) was electro-transformed with 200 ng of pEXP230-cry1Ac-t.After phenotypic expression in JM medium for 3 h (140 rpm and 30°C), serial dilutions of the broth culture were prepared and transformed cells were identified on the basis of colony formation on JM plates containing 40 µg/ml Sm.Transformation was confirmed by AGE and PCR using the primer set described above, with purified plasmid as template.The PAl5 strain carrying pEXP230-cry1Ac-t is hereafter referred to as PAl5(pEXP230-cry1Ac-t).Proteins in PAl5 and PAl5(pEXP230-cry1Ac-t) cell pellets were analysed by SDS-PAGE as described above for the E. coli expression studies.The protein concentrations of samples were estimated by using a Micro BCA Protein Assay Reagent Kit (Pierce, Rockford, USA).Gels were stained by silver staining (Davis et al., 1986).

Toxicity bioassays
E. saccharina eggs were obtained from the South African Sugarcane Research Institute (Mt Edgecombe, South Africa).
Insects were reared on a chickpea diet at 26 ± 2°C, 70 ± 4% relative humidity, and a 12:12 h (L:D) photoperiod.Neonate larvae were collected from egg hatching bags and used directly for neonate-based assays or were reared until they had reached the second instar.After adding an autoclaved agar suspension to the dry ingredients of the chickpea diet, the diet was allowed to cool to 56°C before lyophilized PAl5(pEXP230-cry1Ac-t) or PAl5 was mixed thoroughly with the medium.A multiple-concentration (0.5, 1.0, 2.0, 4.0, and 8.0 mg/ml) bioassay design was used.Bioassays were performed on both neonate and second instar larvae, with five replicates of 20 larvae per concentration.Mortality in each group was recorded 4 d after exposure and corrected for control (chickpea diet only) mortality using Abbott's formula (Abbott, 1925).Larvae that did not move when prodded were recorded as dead.Second instar larvae that survived treatment were individually weighed upon completion of the bioassays.ANOVA was use to analyse the bioassay data.When F values were significant, the Tukey HSD test was used to analyse differences in means (Statistica 7.1, Statsoft, Tulsa, USA).Percentage mortality was transformed by Anscombe's transformation prior to ANOVA (Anscombe, 1948;Zar, 1999).

Protein expression in E. coli
Analysis of the protein profiles of induced BL21-SI(pEXP14-cry1Ac-t) cells showed a novel (not present in BL21-SI cells) protein with an SDS-PAGE-estimated size of 72 kDa (Figure 1).Since cry1Ac-t codes for 636 amino acids with a theoretical molecular weight of 71.1 kDa (estimated with DNAMAN), the size of the novel 72 kDa protein is consistent with that expected from cloned cry1Ac-t.The SDS-PAGE therefore confirmed the functionality of both the cry1Ac-t construct and pEXP14-cry1Ac-t.

Expression of cry1Ac in G. diazotrophicus PAl5
Electroporation was used to transform PAl5 with pEXP230-cry1Ac-t.To confirm the presence of cry1Ac-t in transformed PAl5, plasmids were purified from transformed PAl5 and used as the template for PCR.The PCR resulted in a 2.0 kb product (data not shown), a product consistent with cry1Ac-t.The size of the extracted plasmid (13.9 kb) and the results of the PCR confirmed that PAl5 had been successfully transformed with pEXP230-cry1Ac-t.SDS-PAGE analysis of the recombinant strain PAl5(pEXP230-cry1Ac-t) confirmed the presence of a protein with a molecular weight of 72 kDa (Figure 2).This size is consistent with the Although statistical analyses were performed on transformed data [Anscombe's transformation for percentage (Anscombe, 1948;Zar, 1999)], for clarity the untransformed values are presented in the graph.X-axis is presented on a log base 2 scale.
polypeptide expressed from pEXP14-cry1Ac-t in BL21-SI cells (Figure 1) and with the theoretical molecular weight of the truncated Cry1Ac encoded by cry1Ac-t.

Toxicity bioassays
The toxicity of recombinant G. diazotrophicus PAl5(pEXP230-cry1Ac-t) to neonate larvae was compared to that of non-recombinant PAl5.At both 4.0 and 8.0 mg/ml lyophilized cells, corresponding to 9.1×10 5 and 1.8×10 6 cells/ml respectively, the mortality caused by PAl5(pEXP230-cry1Ac-t) was significantly higher (P<0.05)than that caused by PAl5 (Figure 3).For PAl5(pEXP230-cry1Ac-t), E. saccharina mortality at 4.0 or 8.0 mg/ml was significantly higher (P<0.05)than that recorded at the other concentrations evaluated (Figure 3).In contrast, no significant difference in neonate mortality was observed between different concentrations of PAl5 (Figure 3).Because PAl5(pEXP230-cry1Ac-t) did not kill second instar E. saccharina larvae, the effect of PAl5(pEXP230-cry1Ac-t) on 4-days-post-treatment weight was taken as a measure of toxicity.The interaction between treatment (PAl5(pEXP230-cry1Ac-t) or PAl5) and concentration had a significant effect (ANOVA, F 4.888 =34.2, P<0.00001) on larval weight.At every concentration evaluated, the weights of larvae reared on diet containing PAl5(pEXP230-cry1Ac-t) were significantly lower (P<0.05)than those of larvae reared on diet containing PAl5 (Table 1).In contrast to the results obtained for PAl5(pEXP230-cry1Ac-t), there was a general trend of higher larval weight at higher concentrations of PAl5 (Table 1).

DISCUSSION
With the aim of evaluating the potential of controlling E. saccharina larvae with a G. diazotrophicus strain expressing a Bt insecticidal Cry protein, a truncated cry1Ac gene was cloned into G. diazotrophicus PAl5.Since only the first 612 amino acids of the 1178 amino acid Cry1Ac protein are required to produce an active toxin (Adang et al., 1985), a truncated cry1Ac gene was used in this study.To determine if the truncated cry1Ac gene construct (cry1Ac-t) would be expressed in Gramnegative bacteria, the construct was tested in E. coli using Gateway expression vectors.The production in  a Mean ± standard error; values in a column followed by the same letter are not significantly different (P>0.05;Tukey HSD test for unequal sample size).At each concentration, the mean weight of insects reared on a diet containing PAl5(pEXP230-cry1Ac-t) was significantly lower (P<0.05;Tukey HSD test for unequal sample size) than that of insects reared on a diet containing PAl5.Values in brackets are the sample sizes.
BL21-SI(pEXP14-cry1Ac-t) of a protein with the size expected (71 kDa) from cry1Ac-t confirmed the functionality of the cry1Ac-t construct.Since pENTR-cry1Ac-t was used to produce pEXP14-cry1Ac-t, and the functionality of cry1Ac-t was confirmed in pEXP14-cry1Ac-t, pENTR-cry1Ac-t was used to provide the truncated cry1Ac gene construct for cloning into pKT230.By cloning the cry1Ac-t construct contained in pENTR-cry1Ac-t into pDESTN230 (pKT230 converted into a Gateway-compatible vector) by λ-mediated site-specific cloning, a construct of known functionality was cloned in the desired orientation and reading frame into the vector pDESTN230.The resultant plasmid, pEXP230-cry1Ac-t, was used to transform PAl5.PAl5(pEXP230-cry1Ac-t) was significantly more toxic to neonate larvae than PAl5, with 4.0 and 8.0 mg/ml lyophilized cells of PAl5(pEXP230-cry1Ac-t) causing 4days-post-treatment average mortalities of 20.3 and 29.0% respectively.In comparison, a cry1Ac-expressing recombinant strain Nal1(pMT7) of the obligate endophytic bacterium Herbaspirillum seropedicae caused 48% average mortality 4-days-post-treatment when assayed at 3 mg/ml lyophilized cells against E. saccharina neonate larvae (Downing et al., 2000).However, in contrast to PAl5, non-recombinant H. seropedicae strain Nal1 caused considerable neonate mortality (12%), indicating that the control-corrected mortality for the recombinant strain Nal1(pMT7) was approximately 36% in the study of Downing et al. (2000).
Sublethal effects (feeding inhibition and growth inhibition) are considered more sensitive indicators of Cry protein toxicity than lethal effects (van Frankenhuyzen, 2009).Larvae that survived after treatment with PAl5(pEXP230-cry1Ac-t) showed significant stunting in their growth compared to larvae treated with PAl5.At each concentration evaluated, the weights of second instar larvae feeding on PAl5(pEXP230-cry1Ac-t) were statistically lower than those of second instar larvae feeding on PAl5.Examination of larvae and diet plugs confirmed that the failure to gain weight was due in part to cessation of feeding.Although they were too small to weigh successfully, neonate larvae that survived appeared to stop feeding and did not grow visibly when provided with diet containing more than 1.0 mg/ml PAl5(pEXP230-cry1Ac-t).In contrast, neonates that were fed with diets containing PAl5 continued to feed and developed rapidly.These findings are important because from a pest control perspective cessation of feeding is highly beneficial.
The expression of cry1Ac-t in PAl5 was relatively low, especially when compared to the expression obtained in E. coli, and may have been related to the vector pKT230 and the Km r promoter used.However, previous studies have shown that pKT230 is suitable for the expression of promoterless heterologous genes in G. diazotrophicus and for the expression of cry genes in plant-associated Gram-negative bacteria (Sevilla and Kennedy, 2000;Skøt et al., 1990).Sevilla and Kennedy (2000) used pKT230 to express a promoterless 3.8 kb uidA:cat gene cassette in E. coli and PAl5.The uidA:cat gene cassette was cloned in the SmaI site of the Km r gene.Expression was detected in E. coli and PAl5, however, expression in PAl5 was estimated to be about 50% less than that in E. coli (Sevilla and Kennedy, 2000).Their results confirmed the general suitability of expressing promoterless genes under control of the pKT230 Km r promoter in G. diazotrophicus.However, the expression in PAl5 of a gfp gene under control of the pKT230 Km r promoter was about 90% less than in E. coli (Sevilla and Kennedy, 2000).Poor gfp expression was also obtained with PAl5 carrying pTB93F (Sevilla and Kennedy, 2000), a gfpcarrying plasmid that was successfully used in Rhizobium meliloti (Gage et al., 1996).The results of Sevilla and Kennedy (2000) suggest that the characteristics of the heterologous gene may have a significant effect on expression levels in PAl5.
The average G+C content of PAl5 is 66% (Bertalan et al., 2009;Gillis et al., 1989), with the well-characterized nifHDK (nitrogenase) genes of PAl5 having an average G+C content of 61% (Lee et al., 2000 ).In contrast, the G+C content of the truncated cry1Ac gene is 38%.Analysis of the genes that Sevilla and Kennedy (2000) used in their study shows that uidA (GenBank S69414) and gfp (GenBank AF108217) have G+C contents of 52% and 39% respectively.Since both the gfp and cry1Ac-t were expressed poorly in PAl5, there appears to be an association between low G+C content and poor expression in PAl5.In contrast, both gfp and cry1Ac-t were expressed well in E. coli, a host with a G+C content of approximately 51%.Large differences in G+C content suggest differences in codon usage and codon bias.Comparison of the codon usage in cry1Ac-t with that in the uidA and nifD genes of PAl5 shows significant differences.For example, the codon usage frequency for asparagine is: uidA (AAT 25% and AAC 75%), nifD (AAT 30% and AAC 70%), and the truncated cry1Ac (AAT 71% and AAC 29%).Several studies have shown that protein production in bacteria could be increased by replacing low-usage codons with high-usage codons (Mohsen and Vockley, 1995).
Studies have shown that cry expression may be improved by: (1) including a cry gene transcription terminator (which increases mRNA half-life by protecting the mRNAs from exonucleolytic degradation) (Wong and Chang, 1986), and (2) coexpression of Cry proteins with P20, an accessory protein from Bacillus thuringiensis subsp.israelensis which has been shown to stabilise some Cry proteins (Khasdan et al., 2007;Rang et al., 1996;Shao and Yu, 2004).In addition to testing whether expression of cry1Ac-t in PAl5 may be improved by replacing cry1Ac codons with G. diazotrophicus-preferred codons, further development of the strategy presented in this study should evaluate the effects of a cry gene transcription terminator and the p20 gene on the insecticidal activity of recombinant PAl5 strains.Increased expression of a cry gene in an endophytic bacterium does not necessarily mean that the recombinant bacterium has better pest control potential.For example, high levels of constitutively expressed cry1Ac caused plasmid instability and no detectable protein expression in both H. seropedicae and P. fluorescens (Downing et al., 2000).An endophyte that is sensitive to high levels of Cry1Ac is unlikely to compete with non-recombinant endophytes in planta.A fine balance must be achieved between cry1Ac expression levels, toxicity to the host, and toxicity to the target pest.
In conclusion, recombinant G. diazotrophicus PAl5 carrying a truncated cry1Ac gene from Bt was toxic to E. saccharina larvae.In theory, inoculation of sugarcane stems with the recombinant G. diazotrophicus PAl5 strain would result in Cry1Ac protein being produced in the feeding zones of burrowing larvae and thus reduce damage caused by E. saccharina.Further studies, such as evaluation of the colonization of sugarcane by recombinant G. diazotrophicus strains, are required to determine the feasibility of controlling E. saccharina in the field with recombinant G. diazotrophicus strains Rapulana and Bouwer 1213 expressing a truncated cry1Ac gene.

Figure 1 .
Figure 1.SDS-PAGE analysis of the expression of truncated Bt Cry1Ac protein in BL21-SI cells.An 8% (w/v) polyacrylamide gel and Coomassie Blue staining were used.Lane 1, protein profile of BL21-SI (control); lane 2, protein profile of BL21-SI(pEXP14-cry1Ac-t), with the arrow indicating the position of the truncated Cry1Ac protein (72 kDa).Arrow and number combinations alongside the gel show the positions and sizes (kDa) of the molecular weight markers.BL21-SI samples were prepared for SDS-PAGE by harvesting the cells from 1 ml broth culture, freezing the pellet overnight and re-suspending the cells in 100 µl ×1 loading buffer.For each BL21-SI sample, 20 µl was loaded per lane.

Figure 2 .
Figure 2. SDS-PAGE analysis of the expression of the truncated Bt Cry1Ac protein in G. diazotrophicus PAl5(pEXP230-cry1Ac-t).An 8% (w/v) polyacrylamide gel and silver staining were used.Lane 1, protein profile of G. diazotrophicus PAl5; lane 2, protein profile of G. diazotrophicus PAl5(pEXP230-cry1Ac-t), with the arrow indicating the position of the truncated Cry1Ac protein (72 kDa).Arrow and number combinations alongside the gel show the positions and sizes (kDa) of the molecular weight markers.Lanes were loaded with 10 µg total protein.