Skip to main content

Risk evaluation for genetically modified soil microbial inoculants


The objectives are: the development and evaluation of alternative vector systems without antibiotic resistance genes as selectable markers, and of suicide systems for the biological containment of soil microbial inoculants; risk assessment on the release of soil microorganisms altered in symbiotically relevant traits and evaluation of strains, recovered from inoculated fields, for genetic and phenotypic properties that may be altered since inoculation (computer assisted analysis).
First year objectives of the project were:
development of vector and reporter systems; exploit the natural selection method to obtain various Rhizobium and Bradyrhizobium strains defective in thymidilate synthase, evaluate the performance of thy vector in the rhizosphere under greenhouse and field conditions, and construction of inducible reporter systems.
Construction of a suicide system for R meliloti; evaluate the performance of the R meliloti thy in the rhizosphere under greenhouse and field conditions, and genetically manipulate the Lactococcus lactis thy cartridge in preparation for a conditional expression system.
Evaluation of genetically modified inoculum strains; during inoculant manufacturing process and at microcosm and greenhouse level.
Evaluation of natural occurring populations; to use pyrolysis mass spectrometry (PYMS) to look for distances between isolates of Bradyrhizobium japonicum isolated from soybean plants grown in Italy, to study isolates from Italy using deoxyribonucleic acid (DNA) hybridization, to look at populations of Rhizobium in soil differing characteristics and from different geographic characteristics and from different geographic regions, and to look at the effect of factors such as manipulating the water table on survival.

In work to date the thy system has proved highly effective in ensuring stable maintenance of the plasmid under field conditions. This was found to be the case even though the thy mutant strain was found to revert at a rather high frequency. This demonstrates that the pressure for the thy host strain to revert to a wild type. The lac and lac/Hg{r} cassettes provide a tool which is based on a catabolic inducible reporter system without antibiotic resistance markers. The expression levels, comparable or higher than those obtainable with tac promoter, allow unambiguous detection of the genetically modified microbe. The project is continuing.

Rhizobia are widely used in agriculture as live inoculants for leguminous crops. As a consequence these bacteria represent a major class of man controlled environment impacting microorganisms. Effort was aimed at constructing strains bearing selectable phenotypic traits, namely mercury resistance and beta-galactosidase, devising a microcosm experiment to assess the survival of the modified strains involving stability of the introduced markers and lateral gene transfer to differnet soil bacteria.

The mercury resistant genes were obtained form transposon Tn1831, while the promoterless lacZ gene was from Escherichia coli. The latter is driven by a synthetic promoter to which a regulatory structure, namely the lac operator, has been added to allow inducible expression. The presence of the LacIq was also necessary for the functionality of the regulated system. All of these structures have been asembled into a cartridge. This 10 kilobase pair structure is flanked by unique restriction sites to facilitate cloning procedures. The lac operator can be easily removed, leading ot high level constitutive expression of the lacZ gene.

3 different chimeric plasmids have been prepared, based on the lac-mer catridge. They are currently in coli and are being transferred to Rhizobium leguminosarum biovar viciae strain 1003 via conjugation.
For pDG3 the cartridge is inserted in a broad host range vecotr which is a RSF1010 derivative. It is devoid of any antibiotic resistant gene. Bacteria carrying this will have the reporter genes locatedon a plasmid (ie in an unstable form). The pDG4 plasmid is the same as pDG3 except that the lac oeprator has been removed. It will allow to study the effects of a high level expression of the reporter gene on its stability within the host microorganism.
The pIRL1 a pSUP102 derivative and contains a fragment of R leguminosarum deoxyribonucleic acid (DNA) which carries the recA gene sequence. The cartridge is inserted within the coding region thu s inactivating the recA gene. Since this plasmid is not able to replicate in R leguminosarum, the marker genes can be maintained in the host bacterium only if homologous recombination occurs. In this way it wll be possible to have the reporter genes located on the chromosome (ie in a stable form).

The commercial use of microbial inoculants in agriculture can involve the release of large numbers of genetically modified microorganisms into the rhizosphere. Following such environmental release, it is important to be able to monitor the fate of the released microorganisms. It is also desirable to be able to control the persistence and spread of the released microorganisms.

Research is being carried out into the development of suitable marker genes for the unambiguous detection and identification of inoculant strains. An alternative vector system for the introduction and stable maintenance of recombinant genes into microbial inoculant strains has been developed. This vector system is based on the thymidylate synthase (thy) gene, which is essential for de novo deoxyribonucleic acid (DNA) synthesis and therefore for survival of the organism (in the absence of an exogenous supply of thymidine). Attention has been focussed on the thymidylate synthase gene of Rhizobium meliloti. A stable vector system, based on a host strain defective in thymidylate synthase and a broad host range plasmid, which contains a cloned copy of the Lactococcus lactis thy gene, was evaluated. Under greenhouse and field conditions, this vector system ensured stable maintenance of the plasmid throughout the symbiotic life cycle of the Rhizobium inoculant strain.

The survival and persistence of an R meliloti thy mutant has been monitored in greenhouse experiments, in the presence and absence of the alfalfa host plant. Results from these experiments suggest that the inoculum does not persist in soil or rooting solution, but may survive at a reduced level at the root surface of the host plant.

The R meliloti thy gene has been cloned by complementation. The availability of this gene will facilitate genetic manipulation of the thy locus in R meliloti.

The population dynamics of Bradyrhizobium japonicum 902 A released as soil liquid inoculant have been monitored in the presence and absence of its host plant and at low and high water table levels.

The site where it was possible to control the water table level was selected, and the soil traits (physical chemical and biological properties) along with the cropping history were recorded.
Bradyrhizobium japonicum was released at a density of 1.6 E9 colony forming units (CFU) in each of the 20 m{2} plots, in the presence and absence of the host plant. 4 repetitions for every plot were carried out; the water table was recorded daily and raised, when appropriate, up to 7 times. The inoculant strain was counted (MPN counts) at 30 days and 3 months after inoculation. Before inoculation, rhizobia were found to be below the MPN detection limit (ie less than 0.6 cells/g of soil dry weight).

The results indicated that both presence of the host plant and water table level greatly affected the survival on inoculated Bradyrhizobium japonicum during the first 90 days. The survival in the absence of the host plant decreased by 90% with respect to the initial concentration when the water table was maintained at high levels.

There was a detectable interaction between the presence of the legumionous plant and the depth of the water table with respect to the overall survival rate of the inoculated strain.

Bradyrhizobium japonicum inoculants from fields in Italy are being studied using polymerase chain reaction (PCR) analysis and pyrolysis mass spectrometry (PYMS). For PCR analysis low temperatures are used for annealing primers, giving nonhomologous priming to regions of very similar deoxyribonucleic acid (DNA) sequence, (random amplified polymorphic DNA (RAPD), PYMS has the advantage that it is an assessment of similarity based on phenotypic differences, whereas RAPD shows relatedness at the level of genomic DNA. Therefore RAPD is very useful for confirmation that derivatives of an inoculant are present at a site, while PYMS allows one to see whether there are changes in the phenotype from the strain introduced.

A field was inoculated with Azofix in 1986 and 1989, and nodules were isolated from soybeans in 1989. To date 28 isolates have been examined using PYMS and 8 of these by RAPD. No isolate was found to be derived from Azofix by either technique. Tho isolates were shown to be almost indistinguishable from an inoculant called NPPL which has been used elsewhere in Italy. PYMS produced 2 groups of isolates; 1 very similar to NPPL and the other similar, but distinguishable. It is assumed that the latter group represent a subset of the population that has drifted gentically.

In another field inoculated with Azofix and a Nitragin inoculant, none of the isolates were derived from Azofix even though it was the last inoculant used. These results indicate that Azofix survives poorly in these soils, or that it nodulates poorly in competiton with other bradyrhizobia.

An advantage of RAPD is that the pattern of bands on a gel is simply to recognize and thus if the patterns derived from isolates does not match the expected inoculant they can easily be compared with those derived from other strains in a collection which might have been used, as was done in this study. The combination of RAPD and PYMS also has the advantage that it gives an indicated of how closely st rains are related to each other, and hence how much they might have drifted genetically.
In order to be able to make an assessment of the risks involved in releasing genetically modified microorganisms (GMMs) into the environment, it is necessary to have information on the potential for such microorganisms to persist, multiply and spread.

This project will continue work started in the BAP programme to study the population of (genetically modified) rhizobia released into soils in which they are not indigenous. By isolating rhizobia from such soils it is possible to determine population, and hence survival, persistence and spread in different soils. By analysing the chemistry of the soils, cropping history and other relevant characteristics, it is possible to correlate these with the potential for the (genetically modified) rhizobia to persist. This is being done by using various computer based statistical packages.

The different tasks of the project are distributed among partners as follows: the group at the University College Cork (IE) will deal primarily with the development and evaluation of suicide systems for the biological containment of soil microbial inoculants altered in symbiotically relevant traits, and evaluation of alternative vectors without antibiotic resistance genes; the group at the University of Padova (IT) will focus on the development of vectors systems with highly expressed catabolic marker genes, without antibiotic resistance genes as selectable markers, and on the risk assessment of the GMMs released in microcosm and in the field; the group at the HELIGENETICS (IT) will mainly deal with the baseline study on introduced non-modified and genetically modified soil microbial inoculants; the group at the University of Bristol (GB) will focus primarily on collection of data on the populations and trends in relevant computer based retrieval and statistical packages. Isolated strains will be checked fo genetic and phenotypic properties that may be altered since inoculation.

Funding Scheme

CSC - Cost-sharing contracts


Strada Romea 16
35020 Albignasego

Participants (3)

Heligenetics SpA
Via Provinciale 62/A 12
45030 Gaiba Rovigo
Western Road Biomerit
30 Cork
University of Bristol
United Kingdom
Woodland Road
BS8 1UG Bristol