Actinorhizal and legume hosts share the Common Symbiotic Pathway, a signaling pathway at early stages of interaction that leads to the development of the nodule by the host after recognition of the symbiont , in addition to several key gene orthologs involved in the process of nodule development . On the bacterial side the rhizobial signaling molecules that trigger nodulation in many of the legume hosts, called Nod factors, are synthesized by nod genes and have been extensively studied . By contrast, the majority of Frankia genomes do not contain clear sets of nod gene homologs, with the exception of certain members of the cluster II Frankia whose genomes contain homologs of the rhizobial nodABC genes. These nod gene homologs are expressed in symbiosis . Other mechanisms have been found to trigger nodulation in some legumes by rhizobia, including an effector protein injected into the host by a type III secretion system , however genes responsible for these pathways lack homologs in Frankia genomes as well. Signals that elicit responses in host roots, including root-hair curling, an early step in some rhizobial and actinorhizal symbioses , have been detected experimentally in some Frankia in clusters I and III. The structures of these molecules, however, have not been determined . The major advances in knowledge of rhizobial symbioses have been made possible by the development of genetic tools for dissecting the metabolic and regulatory pathways in the microsymbiont-host interactions.

Long et al. originally demonstrated the role of the nod genes in symbiosis by complementing nodulation-deficient mutants of Sinorhizobium meliloti with nod genes encoded on a replicating,dutch bucket for sale broad host-range plasmid. More recently, expression systems based on marker genes in rhizobia have enabled a wider-range of experiments including the tracking of microsymbionts during the host infection process and the identification of regulatory networks involved in symbiotic interactions . However, to date, there has been no stable electroporation-based genetic transformation system for Frankia , and little understanding of what the barriers to stable transformation may be. Thus an electrotransformation system in Frankia, as reported here, has the potential to enable functional inquiries into the diverse mechanisms involved in establishing rootnodule symbiosis and maintaining biological nitrogen fixation in actinorhizal symbioses. These inquiries will, in turn, contribute to a wider understanding of the origins of root nodule symbioses and mechanisms of Frankia and rhizobial interaction with hosts in the NFC. Frankia presents several barriers to transformation. Actinobacteria in general have low rates of homologous recombination due to competition between their homologous recombination pathway and a Non-Homologous End-Joining pathway . Additionally, natural vectors for Frankia transformation are limited as no Frankia phages have been discovered . Finally, its multicellular hyphae require experiments to be performed on hyphal segments rather than individual cells. Despite these barriers, it has been demonstrated that DNA can be electroporated into Frankia cells , providing a potential avenue for genetic transformation. Furthermore, Kucho et al. reported initial success in integrating a non-replicating plasmid into the chromosome of Frankia sp. CcI3. However the recombined plasmid was lost in the following generation, limiting its further use in experiments.

Restriction enzymes pose a barrier to successful transformation in many bacteria due to their role in defense of digesting foreign DNA. In some actinobacteria, the use of unmethylated plasmids has increased transformation efficiency , potentially due to the expression of type IV methyl-directed restriction enzymes. In the majority of bacterial taxa, DNA is methylated during replication by the methyltransferase Dam, to mark parent strands for DNA repair and excision of misincorporated bases from the daughter strand . The majority of actinobacteria, however, lack dam homologs and an investigation of genomes of Streptomyces, Rhodococcus, and Micromonospora spp. found that these genomes were not methylated in the canonical Dam pattern . While actinobacteria have recently been shown to have a unique mismatch repair pathway that does not involve Dam it is not yet known how this pathway utilizes methylation, if at all. In this study we report successful electrotransformation of Frankia alni ACN14a by using an unmethylated plasmid to circumvent the methylation-targeting restriction enzyme encoded in the ACN14a genome. This permitted the maintenance of plasmids derived from the very broad host-range replicating plasmid pIP501 in F. alni cultures for several months of continuous subculture, and implicates the type IV restriction enzyme as the major barrier to transformation. Use of a replicating plasmid also allowed the maintenance of the plasmid in Frankia without depending on its homologous recombination system. Bioinformatic analysis revealed a differential expression of type IV restriction genes between F. alni transcriptomes in symbiosis vs. culture, which has implications for acquisition of horizontally-transferred genes in endophytic environments by Frankia.

Additionally, we show that egfp expressed from a replicating plasmid can be used to label gene expression differences between Frankia cell types during nitrogen fixation, opening up molecular genetics-based experiments, including marker gene expression, on Frankia symbioses in the future. Restriction enzyme genes from all completely sequenced actinobacterial genomes available in the REBASE database were downloaded with their annotations . Restriction enzymes were categorized by enzyme types: I, II, III, or IV. Bacterial transcriptomes used in this study were downloaded from the NCBI GEO database . Transcriptomes analyzed were chosen based on the following criteria: transcriptomes were made from actively growing pure cultures, the organism whose transcriptome was sequenced did not have any genetic manipulations, e.g., mutations or exogenous plasmids, and media used for sequencing did not have any additional experimental compounds added including antibiotics or complex carbon sources. For comparisons between transcriptomes, expression levels for each gene were calculated as the percent of genes with lower expression than the gene of interest in each transcriptome . A plasmid designated as pIGSAF was designed for constitutive expression of egfp and synthesized by ligating a PCR-amplified fragment containing the egfp gene and promoter from plasmid pDiGc to plasmid pSA3 . Plasmid pSA3 is a broad host-range replicating plasmid originally developed as a shuttle vector by ligating pIP501-derivative pGB305 to the E. coli-specific replicating plasmid pACYC184 . The pIP501 origin of replication has been shown to replicate in bacteria of diverse phyla including Firmicutes, from which it was originally isolated , as well as Actinobacteria and Proteobacteria . Plasmid pSA3 contains chloramphenicol and tetracycline resistance genes and an E. coli-specific p15A origin of replication for higher-copy number propagation . To synthesize PCR products for ligation,hydroponic net pots primers were designed using NCBI Primer-BLAST with default settings . All primers used in this study are listed in Table 1. For cloning, primers were designed with linkers adding target restriction sites onto their 50 ends as well as 4–6 additional bases to aid in restriction digestion of the ends. For the synthesis of pIGSAF, primers were designed targeting the egfp coding region as well as the promoter region 200 base pairs upstream, amplifying a fragment 1238 base pairs in length. Products were synthesized by first performing 10 cycles of amplification using the lower annealing temperature corresponding to the binding site on the target DNA and then an additional 30 cycles using the annealing temperature of the full primer including the linker . The egfp fragment was synthesized with SalI and BamHI restriction sites on the 50 and 30 ends, respectively, with primers EGFP_SalI_F and EGFP_BamHI_R. The PCR product and plasmid pSA3 were then digested with both enzymes. Digestion with SalI and BamHI removed a fragment approximately 200 base pairs in size from plasmid pSA3. The remaining 10 kb fragment was purified on a 0.7% agarose gel and extracted. This fragment was then mixed together with the egfp PCR product, denatured by heating for 5 min at 65◦C, cooled on ice to allow binding of base-paired overhangs, and then the fragments were ligated together by incubation with T4 ligase at 16◦C overnight. The resulting ligation was transformed into E. coli DH5α by heat shock . Transformants were selected on LB plates with chloramphenicol at a final concentration of 25 µg/ml. Transformed colonies were inoculated into liquid LB media with chloramphenicol and cultured as described above. Plasmid pIGSAF was then re-extracted from transformed E. coli and its composition was confirmed by digestion with SalI and BamHI.

The total length of the ligated plasmid was confirmed to be approximately 12 kb by gel electrophoresis. To study the differential expression of egfp in response to nitrogen limitation, a second plasmid, pIGSAFnif, was synthesized that permitted the egfp coding-region to be expressed under the control of the nif cluster promoter of F. alni ACN14a . The nif cluster promoter is upstream of the nif genes coding for the subunits of nitrogenase. Nitrogenase activity has been demonstrated in vesicle fractions of Frankia cultures but not in the hyphal fraction , suggesting that egfp expression under control of the nif cluster promoter will be up-regulated in vesicles relative to hyphae. The egfp coding region of pDiGc was amplified without the upstream promoter region using primers GFP_CDS_EcoRI_F and GFP_CDS_SalI_R . These primers added an EcoRI restriction site before the start codon and a SalI site 200 bases downstream of the stop codon. Separately, the 344 bases upstream of the nif nitrogenase cluster in the F. alni ACN14a genome were amplified with an XbaI site upstream and an EcoRI site downstream using primers nif_promoter_XbaI_F and nif_promoter_EcoRI_R . These two PCR products were digested with EcoRI and ligated together as above to produce a fragment 1134 bp in length. The ligation product was then re-amplified by PCR using the nif promoter forward primer and the egfp reverse primer. The amplified ligation product and plasmid pSA3 were then each digested with SalI and XbaI, ligated together, transformed, and selected on chloramphenicol as above. The final size of plasmid pIGSAFnif was approximately 10.8 kb . F. alni cells were grown in culture for 1 week prior to transformation. Hyphae equivalent to approximately 50 µl packed-cell volume were collected then pelleted as above and resuspended in 500 µl of ice-cold sterile deionized water. This wash step was repeated two more times, but after the third round of centrifugation the pellet was re-suspended in 300 µl ice-cold sterile 10% glycerol instead. Hyphae in the cell suspension were then homogenized by passage through a 21G needle twice. Three hundred microliter of the F. alni cell suspension was pipetted into an electroporation cuvette with a 2 mm gap and mixed with 10 µg of plasmid DNA. Cells were electroporated with unmethylated plasmid prepared as described above, or with methylated plasmid extracted from E. coli DH5α as a control. The cuvette was then incubated on ice for 5 min. Electroporation was carried out in a Bio-Rad Gene PulserTM with Pulse Controller at 2.5 kV, 200  resistance, and 25 µF of capacitance. The cuvette was then immediately filled with 1 ml of ice-cold BAPP media. The cuvette was sealed with ParafilmR and incubated overnight without shaking at 28◦C. The following day the F. alni culture was removed from the cuvette and added to 3.5 ml of sterile BAPP media with tetracycline to prevent contamination, but without chloramphenicol, as above, in a glass 25 ml test tube. The culture was incubated at 28◦C without shaking until visible hyphae were observed . At this point F. alni hyphae were homogenized and sub-cultured into fresh media as above. Subculturing was repeated once more 1 week later. The following week the hyphae were sub-cultured again into BAPP media, this time with chloramphenicol added for selection to a final concentration of 25 µg/ml. Chloramphenicol was chosen as the selective antibiotic because all Frankia strains tested by Tisa et al. were susceptible to it. This process was repeated the following week for an additional round of selection. F. alni cells transformed with the methylated plasmid were similarly cultured. The presence of plasmid pIGSAF electroporated into F. alni cultures was then confirmed by two methods: visualizing total DNA extracts from wild-type and transformed ACN14a on gels and with PCR. For DNA extract visualization, 1 µg each of DNA extract from transformed and untransformed F. alni cultures was loaded into a well on a 0.7% agarose gel. Purified plasmid pIGSAF, extracted with a Qiaprep kit, was loaded into a separate well for comparison.