Segregation pattern of kanamycin resistance marker in Azotobacter vinelandii did not show the constraints expected in a polyploid bacterium

Abstract

It was suggested that Azotobacter vinelandii cells contain about 80 copies of their chromosome and when foreign genes are introduced into the cell, it took several generations for them to spread to all 80 chromosomes even in the presence of selection. In contrast, the fact that many recessive mutants can be isolated from A. vinelandii without the constraints expected for a cell that has 80 copies of its chromosome argued against this organism being highly polyploid. We have investigated the segregation of a kanamycin resistant genetic marker under non-selective conditions in A. vinelandii. Plasmid DNA was used to introduce the kanamycin resistance gene onto the A. vinelandii chromosome at the nifY locus by homologous recombination. The transformants were identified from non-transformants with the aid of replica plating, and hence the colonies examined for segregation of the genetic marker were never subjected to kanamycin selection. In spite of growing the transformants in the absence of selection pressure, no segregant that lacked the kanamycin resistance gene was scored. These analyses suggested that the segregation of the kanamycin marker in A. vinelandii did not exhibit any constraints expected in a highly polyploid bacterium.

Azotobacter, Polyploidy, Segregation pattern, Kanamycin resistance

1 Introduction

Azotobacter vinelandii is a Gram-negative, large soil bacterium capable of fixing nitrogen under strictly aerobic conditions [1, 2]. Its ability to fix atmospheric nitrogen is due to its unique feature of carrying three different, and genetically distinct, nitrogenase systems on its chromosome [3]. Apart from this unique feature, it was speculated that this bacterium contains multiple chromosomes per cell, a situation which can be described as polyploidy [4]. It was reported that exponentially growing A. vinelandii cells contain about 15×10⁻¹⁴ g of DNA per cell [4, 5]. The mass of the Escherichia coli chromosome, on the other hand, is known to be about 3.5×10⁻¹⁵ g per cell [6, 7]. The denaturation and renaturation kinetics experiments, as well as pulse field gel electrophoretic analyses of the A. vinelandii genome after restriction enzyme digestion demonstrated that the size of this chromosome is approximately 4.7 Mbp, which is very similar to the size of the E. coli chromosome [8, 9]. These experiments suggested that A. vinelandii may contain about 40 copies of its chromosome [5, 10]. Moreover, determination of the copy number of the A. vinelandii specific genes nifH, nifD and nifK and foreign genes introduced into the A. vinelandii chromosome, leuB and bla, showed that about 80 copies of these genes are present in these cells [10]. This experiment was done as follows. Assuming that A. vinelandii contained 1.6×10⁵ kb DNA per cell, synthetic mixtures of calf thymus DNA and a plasmid carrying the appropriate gene (native or foreign) were made to generate situations that mimicked the plasmid copy number to be ranging from one copy to 80 copies. Using the appropriate gene as a probe, the intensity of the hybridization signal obtained for the A. vinelandii chromosome and that obtained for different synthetic mixtures were compared. It was observed that the intensity of the hybridization signal obtained for the A. vinelandii chromosome was comparable to the synthetic mixture that mimicked a situation where the plasmid copy number would be up to 80. When a foreign gene such as bla was introduced into the A. vinelandii chromosome by a single point cross-over and DNA was isolated after only a few generations of growth in the presence of ampicillin, it was observed that the intensity of the hybridization signal for the bla probe obtained for such a sample of the A. vinelandii chromosome was comparable to the synthetic mixture that mimicked a situation where the plasmid copy number was less than 80. These observations suggested that several generations are required for the foreign gene (bla) introduced into the A. vinelandii chromosome to spread to all other chromosomes even in the presence of selection (ampicillin). In summary, all these observations suggested that A. vinelandii contains about 40–80 copies of its chromosome and therefore is a highly polyploid organism. Since prokaryotes are normally haploid, these observations imply that A. vinelandii has unique genetics to be investigated and an unusual regulation of its chromosome copy number.

However, some evidence against the natural existence of polyploidy in A. vinelandii cells also comes from genetic analyses of these cells [11, 12]. For example, many mutant strains of A. vinelandii exhibiting the phenotype of a recessive mutation have been isolated [2, 11, 13–18]. These include naturally existing mutants and Nif⁻ mutants isolated by chemical, UV, site-directed or transposon mutagenesis as well as several classes of auxotrophs. These observations suggested that some genetic operations can be performed in A. vinelandii without the constraints expected in a polyploid bacterium. We have used the segregation pattern of a kanamycin resistance marker under non-selective growth conditions as a method to further probe the ploidy status of A. vinelandii. Here we report our genetic analysis on the segregation of a kanamycin resistant gene that was integrated into the unique nifY locus present on the chromosome of A. vinelandii.

2 Materials and methods

2.1 Strains, plasmids and media

The bacterial strain A. vinelandii ATCC12837 used in this study was grown at 30°C in modified Burk nitrogen-free (BN⁻) medium [19]. The BN⁺ medium (pH 7.8) contains 20 g of sucrose; 0.64 g of K₂HPO₄·3H₂O; 0.16 g of KH₂PO₄; 0.142 g of Na₂SO₄; 0.203 g of MgCl₂·6H₂O; 0.074 g of CaCl₂·2H₂O, 0.034 g of FeC₆H₅O₇ and 0.4 g of ammonium acetate per liter. E. coli strains were grown at 37°C in Luria broth, 2YT or minimal medium supplemented with glucose (0.2%) and 10 μg ml⁻¹ of thiamine [20, 21]. When selection was made for the kanamycin resistance marker, the medium was supplemented with kanamycin at a concentration of 25 μg ml⁻¹ for E. coli and 1 μg ml⁻¹ for A. vinelandii. Restriction endonucleases and DNA modification enzymes were purchased either from Boehringer Mannheim (Indianapolis, IN) or from Promega (Madison, WI) and were used according to the instructions outlined in the manufacturers' protocols or following standard methods [20, 21]. DNA fragment purifications were performed using Gene Clean II (BIO 101) following separation on 0.8% agarose TAE preparative gels. Plasmid isolation from E. coli and transformation of E. coli were performed by standard procedures [20, 21].

2.2 Preparation of competent cells of Azotobacter

A. vinelandii cells were made competent according to the procedure described previously [22]. In brief, A. vinelandii was cultured on BN⁺ Fe⁻ agar for about 36 h at 30°C. These iron starved cells were allowed to grow in 10 ml of BN⁺ Fe⁻ liquid medium for about 20 h at 30°C by shaking at 200 rpm. This culture was diluted (1:10) in a total volume of 100 ml of BN⁺ Fe⁻ liquid medium and allowed to grow for 21 h at 30°C by shaking at 200 rpm. Cells prepared in this way were transformed using highly purified plasmid DNA samples as described previously [22] with the modifications described in Section 3.

2.3 Construction of A. vinelandii strain carrying a kanamycin resistance gene on the chromosome at the nifY locus

Initially, we purified a 3.5-kb HindIII restriction fragment carrying nifKTY from pDB6 [13] by separating the DNA fragments on 0.8% agarose TAE preparative gels. The agarose block carrying the 3.5-kb HindIII restriction fragment was excised and the DNA was recovered by using the Gene Clean II system. This fragment was ligated into pUC18 which was predigested with HindIII. The recombinant colonies were identified by a Lac⁻ phenotype exhibited on rich medium supplemented with ampicillin, IPTG and X-gal. One of the plasmids that carry the 3.5-kb HindIII fragment was identified by restriction analysis and designated pBG100. This 3.5-kb fragment contained a site for restriction enzyme BglII that is located in the coding sequence of nifY. To construct pBG102, we purified a 1.1-kb BamHI fragment corresponding to the kanamycin resistance gene from pBG101 and ligated it into pBG100 previously digested with BglII (Fig. 1).

1

Physical and genetic map of pBG102, a plasmid that carries the kanamycin resistance marker inserted into the nifY coding sequence of A. vinelandii. To achieve this, we took advantage of the unique BglII site located in the nifY coding sequence. The DNA corresponding to the kanamycin resistance marker is flanked by the nifY sequence to facilitate homologous recombination.

Our next step was to transfer this inactivated nifY gene (inactivated by the insertion of the kanamycin resistance gene) from pBG102 to the wild-type nifY locus on the A. vinelandii chromosome. Since pBG102 is a derivative of pUC18, this plasmid cannot replicate in A. vinelandii. However, A. vinelandii is known to have a very efficient homologous recombination system. Therefore, it was expected that homologous recombination between the inactivated nifY on pBG102 and the wild-type nifY locus on one of the A. vinelandii chromosomes would occur resulting in the rescue of the nifY::Kan onto the chromosome. Initially, A. vinelandii was transformed with pBG102 to determine whether the kanamycin resistance marker could be rescued onto the A. vinelandii chromosome by this procedure. This experiment showed that Kan^rA. vinelandii transformants could be obtained by this procedure. These Kan^rA. vinelandii transformants were unable to grow on BN⁺ agar plates or BN⁺ liquid medium supplemented with ampicillin. Since pBG102 can confer resistance to both kanamycin and ampicillin, the cells containing this plasmid must be resistant to both these antibiotics. The observation that Kan^rA. vinelandii transformants were sensitive to ampicillin supported the notion that pBG102 was lost from these cells and they had become Kan^r due to the rescue of the nifY::Kan onto the chromosome by homologous recombination.

3 Results and discussion

3.1 Segregation analyses of kanamycin resistance in A. vinelandii strain

Since it was suggested that Azotobacter contained multiple chromosomes [5, 9, 10, 23], we decided to analyze the segregation pattern of the kanamycin resistance marker in order to investigate the number of copies of the chromosome in this organism. The logic behind this experiment is as follows. If A. vinelandii contained only a single genetically active chromosome, all the cells derived from one such A. vinelandii cell would carry the kanamycin resistance marker on the chromosome and would confer kanamycin resistance. If A. vinelandii contained multiple genetically active chromosomes, only some of the cells derived from one such A. vinelandii cell would be carrying the kanamycin resistance marker on the chromosome and would confer kanamycin resistance. Thus, the segregation pattern of kanamycin resistance would reflect the polyploid nature of A. vinelandii cells. To do this, we used the approach described in Fig. 2. The salient feature of this approach is that we did not introduce any selective pressure on the cells. In the transformation reactions, 10–100 μl of competent cells were incubated with different amounts of plasmid DNA ranging from 10 ng to 1 μg. After addition of DNA and incubation at 30°C in MOPS buffer (20 mM MOPS and 16 mM MgSO₄) for 30 min, varying volumes of cell suspension were plated on BN agar plates to obtain single colonies. The ability of these colonies to confer kanamycin resistance was determined by growing them on BN plates supplemented with 1 μg ml⁻¹ of kanamycin, a concentration at which the wild-type A. vinelandii cells could not grow. Simultaneously a replica plate was also made on BN agar (without kanamycin) to maintain this population of colonies without subjecting them to kanamycin selection. Once we identified the colonies that confer kanamycin resistance by observing the growth on BN agar+kanamycin plates, the corresponding colonies from the replica plate (BN agar alone) were re-streaked on new BN agar plates to obtain single colonies as shown in Fig. 2. Twenty-five such colonies were re-streaked on BN agar plates to obtain single colonies. These single colonies were then tested for their ability to grow on BN agar+kanamycin plates. From each original colony (from the above group of 25), 20 single colonies were isolated and tested by this method. Since these colonies originated from a population of bacteria that had never been exposed to kanamycin (or never subjected to selection pressure), we expected that only some of these colonies could receive the chromosome that contained the kanamycin resistance marker. However, all 20 colonies derived from each single colony belonging to the group of 25 original colonies (a total of 500 colonies) that we examined showed growth on BN agar+kanamycin plates. From these 500 colonies, 50 were also maintained on BN agar plates and 10 single colonies were isolated to further verify the possibility of segregation of the kanamycin resistance marker in the absence of selection pressure. However, once again all the single colonies (a total of 500 colonies) showed growth on BN agar+kanamycin plates. This indicates that even in the absence of selection pressure, all the chromosome(s) received the kanamycin marker and all the cells that divided were capable of exhibiting the kanamycin resistant phenotype.

2

The scheme used for examining the segregation pattern of a kanamycin resistance marker integrated into the nifY locus of the chromosome of A. vinelandii. According to this scheme, no selection pressure was exerted on the colonies on which the analysis of the segregation pattern was performed. This was achieved by plating the bacteria after transformation on non-selective BN agar plates and using replica plating as a method to identify the colonies that had received the kanamycin resistance marker on the chromosome (since pBG102 cannot replicate in A. vinelandii, the colonies can become resistant to kanamycin only if the marker from the plasmid is rescued onto the chromosome). The corresponding colonies from BN agar plates were further analyzed by re-growing on BN agar plates to obtain single colonies. If all the daughter colonies exhibit kanamycin resistance in spite of segregation under non-selective conditions, it can be concluded that the cells exhibit characteristics associated with haploid bacteria. In contrast, if only a few of the daughter colonies exhibit kanamycin resistance, this would suggest that the bacterium is polyploid in nature.

3.2 Conclusion

Since the reported DNA content of A. vinelandii cells is much higher than the DNA content of E. coli cells, it was hypothesized that the A. vinelandii cells should have a much larger cell volume to accommodate this large amount of DNA [1, 24, 25]. To test that idea, previously we determined the cell volumes of A. vinelandii using scanning electron microscopy. Our calculations on A. vinelandii cells showed that the cell volume is 12.5 times greater than that of E. coli[1]. Thus, the A. vinelandii cells do have a larger cell volume and therefore could be capable of accommodating a larger amount of nucleic acids.

In this study, we inserted the kanamycin resistance marker onto a unique site in the chromosome of A. vinelandii via homologous recombination using pBG102. Initially, the resulting colonies were grown on non-selective medium after the transformation reaction and the recombinant colonies were identified by replica plating. By following the kanamycin resistance marker and with the help of replica plating, we followed the segregation pattern of this marker. We were unable to identify any segregant that did not carry the kanamycin resistance marker even in the absence of selection pressure. These results show that A. vinelandii does not exhibit the segregation pattern expected of a polyploid bacterium, but instead exhibits a segregation pattern that resembles that of a haploid bacterium (Fig. 2). Thus, while our studies on the cell volume of A. vinelandii indicate the potential of these cells to accommodate a large amount of nucleic acids, the results described here argue against this bacterium being polyploid in nature. If indeed A. vinelandii does contain multiple chromosomes, then to explain the data presented here, we will have to assume that all the chromosomes of this bacterium are not genetically active. Alternatively, the highly efficient recombination system of this bacterium may play a role in creating this genetic uniformity even under non-selective conditions.

Acknowledgements

We thank the members of Gavini and Pulakat laboratories at BGSU for their helpful discussions. This work is supported by Research Challenge Grant, Faculty Research Committee Grant to L.P. and N.G.