Agrobacterium tumefaciens-Induced Bacteraemia Does Not Lead to Reporter Gene Expression in Mouse Organs

Agrobacterium tumefaciens is the main plant biotechnology gene transfer tool with host range which can be extended to non-plant eukaryotic organisms under laboratory conditions. Known medical cases of Agrobacterium species isolation from bloodstream infections necessitate the assessment of biosafety-related risks of A. tumefaciens encounters with mammalian organisms. Here, we studied the survival of A. tumefaciens in bloodstream of mice injected with bacterial cultures. Bacterial titers of 108 CFU were detected in the blood of the injected animals up to two weeks after intravenous injection. Agrobacteria carrying Cauliflower mosaic virus (CaMV) 35S promoter-based constructs and isolated from the injected mice retained their capacity to promote green fluorescent protein (GFP) synthesis in Nicotiana benthamiana leaves. To examine whether or not the injected agrobacteria are able to express in mouse organs, we used an intron-containing GFP (GFPi) reporter driven either by a cytomegalovirus (CMV) promoter or by a CaMV 35S promoter. Western and northern blot analyses as well as RT-PCR analysis of liver, spleen and lung of mice injected with A. tumefaciens detected neither GFP protein nor its transcripts. Thus, bacteraemia induced in mice by A. tumefaciens does not lead to detectible levels of genetic transformation of mouse organs.


Introduction
Genetically engineered plants often represent a preferred source of recombinant proteins and biopharmaceuticals for human consumption [1,2]. In modern plant biotechnology, genetic transformation of plants is usually achieved using Agrobacterium tumefaciens. This bacterium is a soil-borne, nonpathogenic for humans microorganism which can transfer its T-DNA into the genomes not only of plants but also of human cultured cells (for review, see [3][4][5]). Wide exploitation of Agrobacterium for biotechnological purposes and numerous medical cases of Agrobacterium species isolation from bloodstream infections [6][7][8][9][10][11][12][13][14][15][16][17] require the assessment of biosafety-related implications of Agrobacterium invasion of mammalian organisms.
We studied whether or not intravenously injected A .tumefaciens can survive in mouse bloodstream and direct expression of its T-DNA within mouse organs. Our data indicate that, although Agrobacterium persisted in the bloodstream for up to two weeks post injection, it failed to express the reporter GFP gene in such diverse organs as spleen, liver and lung. Thus, Agrobacterium induces bacteremia in mice, but does not cause detectible genetic alteration of mouse tissues.

Results and Discussion
To examine whether agrobacteria can express the reporter gene both in animal and plant cells, we created binary constructs in which GFP expression is driven by either the cytomegalovirus (CMV) or CaMV 35S promoters (Fig. 1). We also inserted a small synthetic intron sequence into the GFP open reading frame (GFPi) to avoid intra-bacterial GFP synthesis due to potentially leaky promoter activity.
Agrobacteria carrying these reporter constructs were first characterized for their viability in blood vessel system. To this end, freshly growing A. tumefaciens GV3101 (10 8 CFU) was administered into mouse by tail vein injection, and blood samples were plated on an antibiotic-containing LB agar medium. Table 1 shows that Agrobacterium remained viable in the bloodstream during at least 6 days after injection. A few blood samples yielded bacterial colonies even two weeks after injection (data not shown). Agrobacteria contained within the blood samples retained not only the capacity to growth on antibiotic-containing media, but also directed expression of the GFP gene from the CaMV 35S promoter in leaves Nicotiana benthamiana (Fig. 2).
Next, we studied the expression of the intron-containing GFPi gene in HeLa cells. On Western blots (Fig. 3), the GFP antibody labeled a 31-kDa band in samples transfected either with CMVdirected GFPi or with CMV promoter-directed GFP (lanes 1, 2). The 27-kDa GFP expressed in mammalian cells is known to assume folding that corresponds to an apparent electrophoretic mobility of a 31 kDa protein [18,19]. This GFP-specific signal was not observed in samples transfected with CaMV 35S promoter-based constructs (Fig. 3, lanes 3, 4). Thus, mammalian cells promoted the correct splicing of the Petunia hybrida intron sequence.
Interestingly, our western blot analyses of proteins from different organs of mice injected with bacteria carrying the CMV promoter-drive reporter construct did not revealed 27-31 kDa GFP-specific products, although some protein samples, including those from control, uninjected mice exhibited nonspecific cross-reactivity of anti-GFP antibodies with a 35-kDa double-band (Fig. 4). Consistently, northern blot hybridization ( Fig. 5) and RT-PCR analysis ( Fig. 6) of RNA isolated from different organs of agroinjected mice did not detect any GFPspecific transcripts. These observations indicate that A. tumefaciens can persist in the bloodstream of agroinjected mice for relatively long periods of time, but they are unable to cause detectible levels of genetic transformation of mouse organs.
That A. tumefaciens and related species can infect mammals, including humans, is supported by numerous medical literature [6][7][8][9][10][11][12][13][14][15][16][17]. However, it should be noted that virtually all the reported cases of such infection occur in immunologically compromised patients who often also have suffered a loss of body integrity, such as invasive surgery, catheter insertion, or a major wound, which presumably allow bacterial access to the bloodstream or internal tissues. Thus, most likely, in healthy individuals, the immune system successfully copes with A. tumefaciens which persists in most of our natural environment, e.g., soil, plant roots, etc. Furthermore, that Agrobacterium strains isolated from infected patients often do not contain the Ti plasmid or its elements required for genetic transformation [6][7][8][9][10][11][12][13][14][15][16][17] suggests that (i) the genetic determinants of this opportunistic infectivity likely reside in the bacterial chromosome, rather than in the Ti plasmid, and (ii) unlike infection of plants, infection of mammals by Agrobacterium does not necessitate genetic alteration of the host. This idea is consistent with our observations that A. tumefaciens can persist in the mouse bloodstream without detectible expression of its T-DNA in the host tissues as well as with previous findings that A. tumefaciens genetically transforms cultured mammalian cells inefficiently [20], probably at the levels undetectable in whole animal tissues. Interestingly, embryonic tissues may be more susceptible to transformation which has been recently reported for sea urchin embryos [21].

Plasmids used for agroinjection
The CaMV 35S promoter-based plasmid encoding GFP was described earlier [22]. A. tumefaciens containing in its T-DNA the GFP gene with Petunia hybrida PSK7 gene 74-nt intron-7 sequence (GFPi) was kindly provided by Dr. Y. Gleba. For construction of the CMV promoter-based vectors, GFP or GFPi cDNA was inserted into the pcDNA3.1 plasmid digested with XhoI and      HindIII. The resulting plasmids were digested with MfeI and SphI, filled-in with the Klenow enzyme and ligated into the binary vector pCAMBIA1300 after its digestion with XhoI and HindIII and filling-in with Klenow.

Agroinjection procedure
A. tumefaciens strain GV3101 was transformed with individual binary constructs, grown in LB medium supplemented with rifampicin 50 mg/l at 28uC. Agrobacterium cells from an overnight culture (2 ml) were collected by centrifugation (10 min, 4,5006g), resuspended in saline and adjusted to a final OD 600 of 0.2. The resulting bacterial suspensions were administered into each recipient naive BALB/c mouse intravenously through the tail vein. For determination of bacterial titers, blood samples (50 ml) from tail vein of mice were added into a tube (950 ml) with heparin (6 units/ml) containing LB growth medium and RGK antibiotic mixture [rifampicin (50 mg/ml), gentamicin (25 mg/ml), kanamycin (100 mg/ml)]. A sample (250 ml) of this mixture was plated on an LB-RGK agar medium, and the bacterial titer was calculated as number of colony forming units (CFU).

RNA isolation and Northern blot analysis
Isolation of total RNA was performed using TRI-REAGENT (Molecular Research Center, Inc) using the manufacturer's protocol. Northern blot analysis was performed as described earlier [23].
Reverse Transcription-Polymerase Chain Reaction (RT-PCR) analysis RT of extracted RNA was performed using the first-strand cDNA synthesis kit (Promega) according to the manufacturer's instructions. First-strand GFP and actin cDNAs were synthesized using 0.5 mg of RNA and primers GFP-M (59TTACTTGTA-CAGCTCGTCCATGCCGAGA39) and Act-M (59AGGGTAC-ATGGTGGTGCCGCCAGAC39), respectively. The cDNAs were then amplified by PCR using primers GFP-P (59ATGGT-GAGCAAGGGCGAGGAGCTGTTC39) and Act-P (59CCAAGGCCAACCGCGAGAAGATGAC39), respectively. PCR was carried out in a programmable Tercik thermocycler (DNA technology, Russia) with the following conditions: 94uC for 10 minutes followed by 30 cycles, each comprising denaturation for 1 minute at 94uC; annealing for 1 minute at 58uC for GFP or 60uC for actin; and then extension for 1 minute at 72uC. After completion of PCR, reaction tubes were kept for 5 minutes at 72uC and then stored at 4uC until use. Negative controls routinely used for each set of primers included reactions without template. Samples were analyzed on 1.5% agarose gels.

Western blot analysis
The proteins from mouse organs (spleen, liver, lung) were isolated using TRI-REAGENT (Molecular Research Center, Inc) according to the manufacturer's protocol and analyzed by western blotting with GFP-specific antibodies as described earlier [23].