Results and Milestones 2nd year of the GMOBILITY project

WP 1: Selection, construction and lab testing of donor sequences and recipient strains for HGT

WP1 is completed. For investigation of HGT in food, B. subtilis was equipped with the marker rescue system based on repair of the 10-bp deleted nptII on a broad host range plasmid. The functionality of the marker rescue system in B. subtilis was demonstrated in vitro.

For studying the HGT of mobilizable plasmids a two-plasmid system was constructed: one helper plasmid pIL205 (tra+ mob+ Cmr) and a mobilizable plasmid (luxAB, Emr). Conjugal transfer of the mobilizable plasmid occurred with low frequency in vitro. For monitoring HGT in Gram-negative bacteria, a plasmid pBBR122 del (Cmr, Kms) was constructed that contains a deleted kanamycin resistance gene for marker rescue in Bacteroides fragilis.

WP 2: Horizontal gene transfer in food systems

DNA stability in rumen model was analysed. Both plasmid and GM-potato DNA were rapidly degraded in the rumen flora. Free plasmid DNA lost its capability to transform E. coli; a half-life period of ca. 1 minute was observed. DNA present in grinded GM-potatoes was rapidly degraded under the limit of PCR detection within 2 h. HGT in the rumen system was studied. Streptococcus bovis could only be transformed by plasmid pGKV21 under laboratory conditions. Rumen conditions abolished the transformation potential. Furthermore the S. bovis strain is not an appropriate hosts for marker rescue nptII gene because it is kanamycin resistant (1 mg/ml Km). E. coli, and Streptococcus gordonii are currently used as model bacteria for HGT.

DNA stability in food was analyzed. DNA in purchased soy products remains detectable over a long period of time. Ingredients such as arginine, polyamines, and biogenic amines (positively charged molecules) exert protective effects on DNA against nuclease attack, whereas ingredients such as glutamate or nitrite did not. During food processing conditions, e.g. heat (85°C) and/or low pH (pH 4), maize DNA and plasmid DNA were rapidly degraded. Alkaline cooking during production of instant "masa mahrina" efficiently degraded the maize DNA to small fragments (<500 bp). During heat treatment (100°C, 30 to 60 min) in soymilk production, the PCR detectable DNA concentration decreased. Fragments up to 1.7 kb remain detectable up to 30 min of processing. Upon the production of chips, flakes and dried potato sticks from the GM-potato degradation of the plant DNA was monitored. Apart from DNA degradation, adsorbtion of bacterial DNA to the food matrix decreases the DNA availability for transformation of B. subtilis in food.

HGT in food is under investigation. Competence development of various bacteria in food is an important item for HGT. Bacillus subtilis and Acinetobacter spp. were shown to develop competence for transformation in various food products such as soy milk, chocolate milk, cooked potato extract etc.

Uptake and integration of plasmid DNA into the chromosome via homologous recombination was investigated in B. subtilis. Frequencies of 10-6 were obtained in soymilk. Marker rescue of nptII in B. subtilis with donor DNA of E. coli was observed in milk products with frequencies of 10-8 to 10-6. Marker rescue of nptII in Acinetobacter spp. was observed with donor DNA of E. coli in various foods with frequencies ranging from 10-4 to 10-7. Marker rescue of nptII with genomic DNA of the GM-potato took place on potato slices and in cooked potato extract with frequencies close to the limit of detection (ca. 10-8).

WP 3: Horizontal gene transfer in vitro

Integrity of free DNA was studied in the in vitro gastrointestinal tract model (TIM). DNA stability in the presence of a starch rich diet (spaghetti) and a protein rich diet (milk) was compared. DNA degradation in the stomach compartment proceeds rapidly 45 minutes after ingestion of both starch and protein rich meal. Free DNA is most rapidly degraded in the duodenum compartment compared to the other compartments. No significant differences were observed between the speed of DNA degradation of both meals. Stability of the nucleases in the enzymatic mixture of the pancreas in the two different digestions situations are envisaged to be different. Probably there are additional mechanisms that have an influence on DNA degradation. Only a marginal difference in influence of the food matrix on DNA degradation (protein rich versus starch rich meal) was found. To validate the in vitro data, the results are compared with the data of tests in rodents and intestinal samples from pigs, as an accepted model for the human intestinal situation. HGT in vitro. Members of the complex micro flora of the GI tract may take up DNA in a functional way. To analyse this scenario, initial studies showed that E. coli is able to take up DNA in environments mimicking the gastrointestinal situation, such as ileum and colon. Transformation of E. coli to ampicillin resistance with plasmid pEMBL8 was only shown at low frequency in ileum juice from a milk digestion run in TIM, when calcium chloride was present or upon heat shock (42°C). In normal ileal juice, the transformation efficiency was below detection level. Transformation of E. coli in colon medium without a bacterial population was observed with low frequency. Calcium chloride enhanced transformation efficiency. The detection of a transformation event was enhanced by selection at sublethal amounts of ampicillin and prolonged incubation of the transformation mixture in colon medium. Monitoring of transformation in the colon flora was difficult due to the overgrowth of too many other bacteria. So far no transformants were detected in the colon flora. Since both E. coli and colon flora degrades DNA within 1 hour to non-detectable amounts, transformation experiments are now conducted by the constant "feeding" of DNA.

WP 4: Horizontal gene transfer in vivo

Detection of DNA stability from feed after ingestion by quantitative PCR methods has been optimized to detect GM maize DNA both in the food material and during digestion through the gut. Only with irradiated maize flour could DNA sequences be detected. This flour was used in further studies to track the stability through the gut. Invertase DNA fragments have been detected in stomach, duodenum and some faecal samples of rats. DNA sequences from plasmid DNA in gnotobiotic rats associated with either E. coli or Bacillus subtilis could not be detected in faeces collected 24 hours after dosing. However, DNA was found in the various compartments of the intestine (ileum, duodenum) that enabled transformation of competent bacteria.

In vivo horizontal gene transfer via conjugation in rodent models was established. In these studies a recipient Enterococcus faecalis strain in the intestines of germfree mice acquired the conjugative plasmid pIL205, and/or the mobilizable plasmid, pCAC from the Lactococcus lactis donors. The orientation of the lux:eryR gene-cassette relative to the mob gene in pCAC influenced the efficiency of transfer. The transfer rate for the plasmid pCAC2, carrying the first orientation, was extremely low (only one transfer event detected in many experiments). Studies using the plasmid, pCAC4, in which the gene-cassette was in the second orientation, did indeed demonstrate that orientation to the mob gene was important. Transfer of pCAC4 occurred in the range of 101 to 103 cfu/g faeces. Transfer of plasmid pIL205 was observed in E. faecalis transconjugants with a frequency of 104 cfu/g faeces). In dixenic mice with a flora from human faeces, the E. faecalis recipient was considerably repressed to about 5 log below that seen when only the donor and recipient strains were present.

As part of the evaluation of the used in vivo models, similar studies in gnotobiotic rats confirmed these findings. HGT in rats associated with the E. faecalis recipient and that received single doses of the donor L. lactis carrying either orientations of the plasmid, pCAC2 or pCAC4 was performed. Only transfer of pCAC4 was observed as well as transfers of the conjugative plasmid pIL205 in these rats.

Colonisation of recipient bacteria for transformation in the intestinal tract of rats were studied Three E. coli strains were able to colonise the gut resulting in faecal concentrations of between 108 to 109 cfu/g faeces. In contrast Acinetobacter calcoaceticus was unable to colonise the gut. B. subtilis was able to colonise at low levels (105 cfu/g faeces).

Colonisation of donor and recipient strains was studied in the intestines of mice carrying a conventional gut flora. Colonization of E. coli strains and a Pseudomonas stutzeri strain was shown in the intestine of conventional mice. All strains survived the gut at low levels and retained their plasmids. Studies with E. coli, using both donor and recipient strains however could not demonstrate any plasmid transfer.

The effect of diet on conjugal transfer in the gut is under investigation with L. lactis, carrying pIL205 and pCAC4 as a donor and recipient E. faecalis strain as used in the transfer studies in mice and rats.

WP 5:

Different marker/donor/recipient combinations for further analysis in the different models are under investigation now. These include conjugative plasmid system pIL205/ mobilisable plasmid pCAC2 and pCAC4, from donor L. lactis to recipient E. feacalis and marker rescue system for the antibiotic resistance genes nptII with recipients such as Bacillus subtilis, Acinetobacter, S. gordonii.

Exposure assessment was conducted to estimate the likely daily intake of maize DNA through the diet. This is required for an overall quantification of the risk of horizontal gene transfer (HGT) in the human gut. For the estimation, data of food consumption data of the Dutch National Food Consumption Survey (DNFCS) of 1997 / 1998 was used, which resulted in a top 20 list of food products that contributed most to the intake of maize in the Dutch population. In addition, the maize concentration within these products was estimated. The daily intake of maize was assessed and combined with the amount (and integrity) of maize DNA found in these products.