Allele-specific PCR shows that genetic exchange occurs among genetically diverse Nodularia (Cyanobacteria) filaments in the Baltic Sea

Gary L. A. Barker1, Barbara A. Handley1, Panmuk Vacharapiyasophon1, Jamie R. Stevens2 and Paul K. Hayes1

School of Biological Sciences, University of Bristol, Woodland Road, Bristol BS8 1UG, UK1
Hatherly Laboratories, School of Biological Sciences, University of Exeter, Prince of Wales Road, Exeter EX4 4PS, UK2

Author for correspondence: Paul K. Hayes. Tel: +44 117 928 7483. e-mail: Paul.Hayes{at}Bristol.ac.uk


   ABSTRACT
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Some cyanobacteria have been shown to exchange genetic information under laboratory conditions, but it has not been clear whether such genetic exchange occurs in the natural environment. To address this, a population genetic study was carried out on the filamentous diazotrophic cyanobacterium Nodularia in the Baltic Sea. Nodularia filaments were collected from 20 widely distributed sampling stations in the Baltic Sea during June and July 1998. Allele-specific PCR (AS-PCR) was used to characterize over 2000 filaments at three loci: a non-coding spacer between adjacent copies of the main structural gas vesicle gene gvpA (gvpA-IGS), the phycocyanin intergenic spacer (PC-IGS) and the rDNA internal transcribed spacer (rDNA-ITS). The three loci were all found to be polymorphic in the 1998 population: two alternative alleles were distinguished at the gvpA-IGS and PC-IGS loci, and three at the rDNA-ITS locus. All 12 possible combinations of alleles were found in the filaments studied, but some were much more common than others. The index of association (IA) for all possible pairwise combinations of isolates was found to differ significantly from zero, which implies that there is some linkage disequilibrium between loci. The IA values for 16 out of 20 individual sampling stations also differed significantly from zero: this shows that the observed linkage disequilibrium is not due to pooling data from genetically distinct subpopulations. Monte-Carlo simulations with random subsets of the data confirmed that some combinations showed significantly more linkage disequilibrium than expected by chance alone. It is concluded that genetic exchange occurs in the natural Nodularia population, but the frequency is not high enough for the loci to be in linkage equilibrium. The distribution of the 12 genotypes across the Baltic Sea was found to be non-random, but did not correlate with temperature, salinity or major nutrient concentrations. A significant relationship was found between the gene diversity among filaments at each station and the distance of the station from the centre of the sampling area: possible reasons for this trend are discussed.

Keywords: AS-PCR, index of association, population genetics, gene diversity

Abbreviations: AS-PCR, allele-specific polymerase chain reaction; IGS, intergenic spacer; ITS, internal transcribed spacer; PC, phycocyanin


   INTRODUCTION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Many agronomically and medically important bacteria have been shown to undergo extensive genetic recombination (Maynard Smith et al., 1993 ). Although there have been numerous population genetic studies of pathogens (e.g. Spratt et al., 1995 ; Stevens & Tibayrenc, 1996 ; Whittam, 1995 ; Trott et al., 1997 ; Feizabadi et al., 1997 ) and soil micro-organisms (Eardly et al., 1990 ; Hagen & Hamrick, 1996 ; Silva et al., 1999 ), there are few such studies for aquatic micro-organisms (Wise et al., 1996 ) and none for cyanobacteria, despite the fact that many members of the group form toxic blooms which threaten water quality. Laboratory studies of cyanobacterial cultures have shown that there is the potential for gene exchange in the environment. Conjugation has been demonstrated both within (Muropastor et al., 1994 ) and between genera (Sode et al., 1992 ; Chiang et al., 1992 ; Thiel & Wolk, 1987 ). Many taxa are naturally competent (Singh et al., 1987 ; Barten & Lill, 1995 ), and studies have shown that potentially transforming DNA is often present in seawater (Lorenz & Wackernagel, 1994 ). Transduction is also a potential mode of genetic exchange, as cyanophages are known to be present in high concentrations in most aquatic environments (Wilson & Mann, 1997 ).

The filamentous, diazotrophic cyanobacterium Nodularia is a convenient organism on which to perform population genetic analysis as it is reliably present during the summer months in the Baltic Sea, and sufficient template DNA for PCR can be obtained from single filaments isolated from the water column. Whilst the extraction of DNA from filaments precludes any further analysis of them, the advantages of directly analysing filaments picked from water samples, rather than clonal cultures established from them, are twofold: first, a great deal of time and effort is saved in isolating and maintaining thousands of clonal cultures; and secondly (and more importantly), culture collections may not be representative of the natural population if genetic differences between isolates have differential effects on fitness in the artificial conditions of laboratory culture.

Genetic diversity in natural populations of Nodularia has been demonstrated in the Baltic Sea (Hayes & Baker, 1997 ; Barker et al., 1999 ) and elsewhere (Neilan et al., 1995 ; Bolch et al., 1996 , 1999 ), but these studies have either focussed on a single locus (Hayes & Barker, 1997 ; Neilan et al., 1995 ; Bolch et al., 1999 ) or had a sample size too small for population analysis (Barker et al., 1999 ). In this study we have characterized three loci from 2336 Nodularia filaments collected during research cruises in June and July 1998. The large sample size minimizes the sampling error of population genetic parameters such as gene diversity (Nei, 1987 ) and index of association (IA) (Maynard Smith et al., 1993 ). The three loci chosen were (1) part of the cpc operon that includes the phycocyanin intergenic spacer (PC-IGS; Neilan et al., 1995 ), (2) the rDNA internal transcribed spacer (rDNA-ITS; Wilmotte, 1994 ) and (3) the intergenic spacer between two adjacent copies of gvpA (the gene encoding the main structural gas vesicle protein; Tandeau de Marsac et al., 1985 ), gvpA-IGS. These loci were selected for study because each is known to be polymorphic in Nodularia (Barker et al., 1999 ). For the PC-IGS locus four sequence variants were found among cultured isolates. Sequence analysis of cultured isolates similarly revealed the presence of three distinct rDNA-ITS, and two gvpA-IGS genotypes (Barker et al., 1999 ). The four PC-IGS sequence types fall into two main groups that can be distinguished by allele-specific PCR (AS-PCR) (Hayes & Barker, 1997 ), as can the three rDNA-ITS sequence types. The AS-PCR approach uses PCR to amplify a region with broad-specificity primers and then a second, nested PCR with one broad-specificity and one or more allele-specific primers: the presence/absence and/or size of the resulting PCR products indicate which allele(s) were present in the original template DNA.

In addition to questions regarding population structure, there is currently disagreement concerning the taxonomic status of Nodularia in the Baltic Sea, with some studies suggesting that several species co-exist in this region (Komárek et al., 1993 ; Bolch et al., 1999 ), whilst other data show that many phenotypic and genotypic markers are not congruent, implying that gene flow occurs between isolates (Barker et al., 1999 ). The taxonomic implications of our results are discussed.


   METHODS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Sampling.
Nodularia filaments were collected from the mixed layer of the water column during June and July 1998 using a 100 µm mesh plankton net as described by Barker et al. (1999 ). Sampling locations are shown in Table 1. Individual filaments were picked from ~50 µl drops of population sample, washed, and isolated with the aid of a dissecting microscope (Walsby et al., 1995 ). To avoid sample bias (for example systematically picking long filaments in preference to shorter ones), plankton catches were diluted with filtered sea water to give<5 filaments per 50 µl and then every filament in a drop of the sample was picked for analysis. Washed filaments were transferred to 0·5 ml Eppendorf tubes containing 25 µl Supertaq PCR buffer: 20 mM Tris/HCl (pH 9·0), 3·0 mM MgCl2, 100 mM KCl, 0·2% (w/v) Triton X-100, 0·02% (w/v) gelatin (HT Biotechnology). Tubes were overlaid with light mineral oil (Sigma), heated by immersion in freshly boiled water for 5 min to lyse the cells, and stored at -20 °C: the filament lysate was used, without further treatment, as the source of template DNA in the reactions described below.


View this table:
[in this window]
[in a new window]
 
Table 1. Sampling location and genotype data

 
Multilocus genotype determination using PCR.
An overview of the PCR reactions performed, and the primers used at each stage, is shown in Fig. 1. Amplification of the PC-IGS and rDNA-ITS loci was performed in two stages. In the first stage, the primer pairs PCßF/PC{alpha}R and 1358F/ITSREV3 (Table 2) were used to amplify both the PC-IGS and rDNA-ITS regions in a duplex reaction. In separate second-stage reactions the forward primers PCßF and 1358F were used in combination with allele-specific reverse primers (Fig. 1, Table 2) to reveal the genotype of each filament. Amplification of the gvpA-IGS was carried out in a single stage, and alternative alleles were detected as product length variants. The PC-IGS/rDNA-ITS stage 1 reaction was carried out in a volume of 32 µl; all other reactions were carried out in 25 µl volumes. All amplification reactions contained Supertaq buffer to 1x final concentration, 12·5 pmol of each primer, 100 µM deoxynucleoside triphosphates (Pharmacia), and 0·5 units Supertaq DNA polymerase (HT Biotechnology). Reactions were overlaid with light mineral oil (Sigma), and subjected to the thermocycling parameters shown in Table 3 using an Omnigene thermocycler (Hybaid). Negative control tubes, containing all reaction components except Nodularia filaments, and positive controls, containing single filaments of Nodularia isolate BC-Nod 9402, 9408 or 9427 (Barker et al., 1999 ), were set up for each batch of test filaments, and subjected to the two-stage amplification procedures described above. All reaction products were resolved on agarose gels as described by Hayes & Barker (1997 ).



View larger version (23K):
[in this window]
[in a new window]
 
Fig. 1. Flow diagram showing the amplification reactions used to determine the multilocus genotypes of single Nodularia filaments. The primers (shown in bold type) and template volumes used at each stage are indicated. The PC-IGS/rDNA-ITS stage 1 reaction was carried out in a volume of 32 µl; all other reactions were carried out in 25 µl volumes. The primer sequences are shown in Table 2; other reaction conditions are described in the text and in Table 3.

 

View this table:
[in this window]
[in a new window]
 
Table 2. PCR primers used in this study

 

View this table:
[in this window]
[in a new window]
 
Table 3. Amplification conditions used in this study

 
Detection of linkage disequilibrium.
The index of association (Maynard Smith et al., 1993 ) was used to determine whether the Nodularia population was in linkage equilibrium (panmixis) or disequilibrium (predominantly clonal growth). The index of association IA is equal to (Vobs/Vexp) - 1, where Vobs is the variance of pairwise genetic distance values for all possible combinations of individuals, and Vexp is the variance expected if the population is in linkage equilibrium. Values of IA which differ significantly from zero are indicative of some degree of linkage disequilibrium (Maynard Smith, 1995 ). For the Baltic Sea as a whole, a Monte-Carlo approach was used to test whether the observed IA value was significantly different from that expected if the population was in linkage equilibrium. Five hundred new random data sets were produced with identical allele frequencies at each locus to the original data, and an IA value calculated for each one. The standard error on all 500 IA values was then calculated. For the calculation of IA for individual sampling stations, the expected error was calculated as described by Maynard Smith (1995 ).

The association of alleles at different loci in Nodularia was also studied using the method of Stevens & Tibayrenc (1996 ), a modified Mantel test (Mantel, 1967 ), to detect linkage disequilibrium based on correlations between independent genetic markers (the PC-IGS, rDNA-ITS and gvpA-IGS loci); non-random associations can provide evidence that gene flow is restricted in the population under study, irrespective of the reasons for this. The technique is a generalized Monte-Carlo test for linkage, independent of mating system or ploidy. In brief, the loci are shared into two groups (for example PC-IGS in one group and both rDNA-ITS and gvpA-IGS in the other). Next, a genetic distance matrix is calculated for each of these groups, showing the distance between a given individual and all other members of its group. A correlation coefficient (r) is then calculated between the two groups of distances for this, the ‘observed’ arrangement of genotypes. The genotypes (the allele patterns for each locus for each Nodularia filament) are then randomly reassorted a given number of times (e.g. 100 times), each time creating a ‘new’ set of Nodularia ‘specimens’ of randomly mixed genetic composition. A new r-value is calculated for each random reassortment. Following the required number of genotype reassortments and r-value calculations, a frequency distribution of r-values can be constructed, allowing a probability value to be attached to the observed correlation coefficient. The process is then repeated and the loci are shared into two different groups. When the number of loci is six or less, as in this study, this is repeated until all combinations (three in this study) of loci have been evaluated. For each analysis, 100 random reassortments of each of the three combinations were used to construct each distribution of r-values. Significance levels of P<0·01 and P<0·05 were used for each combination. Computational constraints necessitated that linkage analyses were performed on 20 subsets of 100 randomly selected filaments, rather than the full 2336.

Other statistical analyses.
Gene diversity was calculated for the population as a whole, and for individual sampling stations as described by Wise et al. (1996 ). Genetic distance was calculated as described by Nei (1972 ) with distance D equal to -logeI, where I, the normalized identity, is equal to Jxy/(Jxx Jyy)0·5. If there are three alleles at a given locus in population X with frequencies of p1, p2 and p3, then Jxx={Sigma}p12+p22+p32. If the equivalent allele frequencies in population Y are q1, q2 and q3, then Jyy={Sigma}q12+q22+q32, and Jxy={Sigma}p1q1+p2q2+p3q3. The arithmetic mean of each J value over all loci is used to calculate I. All data manipulations and statistical analyses were performed using Microsoft Excel 8.0 and MINITAB, except for the modified Mantel test, which was performed using DLX.FOR (Stevens & Tibayrenc, 1995 ).


   RESULTS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Using PCR to determine the multilocus genotype of filaments
The PCR products associated with each allele at the three loci studied are shown in Fig. 2. Filaments with PC-IGS AS-PCR products of 336 bp and 526 bp were designated types III (BC-Nod 9402 lane 1) and I/II (BC-Nod 9401 and 9427 lane 1) respectively: the AS-PCR cannot distinguish the closely related sequence types I and II (Hayes & Barker, 1997 ; Barker et al., 1999 ). Filaments able to produce a PCR product with primers 9408ITSR or 9427ITSR were designated rDNA-ITS types I (BC-Nod 9401 lane 2) and II (BC-Nod 9427 lane 3) respectively, and filaments producing a product with both primers were designated type III (BC-Nod 9402 lanes 2 and 3). The latter class produces a PCR product with both 9408ITSR and 9427ITSR because it contains two non-identical copies of the rDNA region (see Barker et al., 1999 , for detailed explanation). Filaments yielding 280 bp and 380 bp gvpA-IGS PCR products were designated gvpA-IGS types I (BC-Nod 9401 and 9427 lane 4) and II (BC-Nod 9402 lane 4) respectively. All AS-PCR reactions yielded products of the expected sizes only. The full-length PCR products of the first-stage amplification of the PC-IGS and rDNA-ITS (Fig. 1) were occasionally present along with the smaller AS-PCR products. These are not visible in Fig. 2, but, where present, were easily resolved from the AS-PCR products.



View larger version (82K):
[in this window]
[in a new window]
 
Fig. 2. Agarose gel showing AS-PCR products for cultured Nodularia isolates BC-Nod 9401, BC-Nod 9402 and BC-Nod 9427 (Barker et al., 1999 ). The four lanes for each isolate are (1) PC-IGS, (2) rDNA-ITS using primer 9408R, (3) rDNA-ITS using primer 9427R and (4) gvpA-IGS. Size marker lanes (M) are loaded with 100 bp ladder: the bright band in the marker lanes has a length of 500 bp.

 
Genetic structure of the population
Five hundred filaments were collected at each of 20 sampling stations across the Baltic Sea (Table 1, Fig. 3), and multilocus genotyping was carried out on a total of 4143 single filaments, 141 to 310 filaments from each sampling location. It was not possible to determine the genotype at all three loci for every filament: 56% (2336) gave results at all three loci and a further 38% (1562) at only one or two loci, with only 6% (245) failing to produce any amplification products (Table 4). The three loci studied yielded results with varying degrees of success: the gvpA-IGS PCR primers gave products for 87% (3601) of filaments analysed, the rDNA-ITS primers for 75% (3122) and the PC-IGS primers for 65% (2690) (Table 4). At each locus a single allele was found to be dominant within the total population: at the gvpA-IGS locus allele I was carried by 97% of the 3601 filaments that produced a PCR product, for the rDNA-ITS locus it was allele II that dominated (92% of the 3122 successful amplifications) and at the PC-IGS the population was dominated by allele III (87% of the 2690 successful amplifications). There are small, but significant differences in the PC-IGS and rDNA-ITS allelic ratios in those filaments that amplified at all three loci and those that amplified at less than three loci (Table 5) (PC-IGS {chi}2=30·52, d.f.=1, P<0·001; rDNA-ITS {chi}2=7·01, d.f.=2, P<0·05; gvpA-IGS {chi}2=0·01, d.f.=1, 0·95>P>0·90). The origin of these differences is uncertain. It is possible that limiting amounts of template DNA were liberated from those filaments where one or more amplifications failed and that under these conditions small differences in primer annealing characteristics resulted in the underdetection of the most abundant allelic variants at the PC-IGS and rDNA-ITS loci. For the following population analysis we only consider the 2336 filaments that gave results at all three loci, where there is a reduced likelihood of any bias in the data.



View larger version (22K):
[in this window]
[in a new window]
 
Fig. 3. Sampling stations used in this study. Alexander von Humboldt and Litorina sampling stations are shown as open and filled boxes respectively. Station 1, where Nodularia was not found is shown as a filled circle. This figure was redrawn from an area map with marked sampling stations created at http://pubweb.parc.xerox.com/

 

View this table:
[in this window]
[in a new window]
 
Table 4. Results from the 4143 PCR reactions carried out in this study

 

View this table:
[in this window]
[in a new window]
 
Table 5. Allelic ratios

 
Spatial distribution
The distribution of the most common genotype, type 9 (Table 1), was examined to see if it was evenly distributed across all sampling stations. Chi-squared analysis was used to evaluate the null hypothesis that the relative abundance of this genotype was the same at each station. The null hypothesis of equal relative abundance was rejected ({chi}2=126, d.f.=19, P<0·005); therefore we conclude that the distribution of this genotype is not random. Neither parametric (Pearson) nor non-parametric (Spearman rank) tests revealed any correlation between either the individual allele frequencies, or the proportion of genotype 9 at each sampling station, and either temperature, salinity, depth or nutrients including nitrate, nitrite, ammonia, phosphate and silica. The remaining 11 genotypes were too scarce for statistical analyses of their spatial distribution to be performed.

The hypothesis that the genetic distance between sampling stations might correlate with the map distance was tested by calculating Nei’s normalized identity I and thus the genetic distance D for each pairwise combination of sampling stations. Using linear regression, we found no significant relationship between these variables (R2=1·2%), but although there is no significant trend, the scatter plot (Fig. 4) shows that at the greatest distances compared, the genetic distance between stations is always low. In order to test the hypothesis that this effect was caused by low genetic diversity at the extreme Northern and Southern limits of the sampling area, the gene diversity, averaged over all three loci, was calculated for each of the 20 stations, and compared to the distance between each station and the exact centre of the area sampled. The results (Fig. 5) show that there is a significant relationship between gene diversity and the logarithm of the distance from the centre of the sampling locations (P= 0·037, R2=22·1%), following the equation gene diversity=0·253-(0·587 log distance).



View larger version (16K):
[in this window]
[in a new window]
 
Fig. 4. Scatter plot and regression line showing the relationship between the genetic and physical distance between all pairwise combinations of sampling stations.

 


View larger version (11K):
[in this window]
[in a new window]
 
Fig. 5. The relationship between the gene diversity at each sampling station, and its distance from the exact centre of all sampling stations.

 
Linkage analysis
Taken as a whole, the population was found to be dominated by filaments of genotype 9 (Table 1), which accounted for 81% of the population and at least 59% at individual stations. All of the 12 possible PC-IGS/rDNA-ITS/gvpA-IGS allele combinations were found to be present in the population, although several were extremely rare (Table 1). The index of association (IA) calculated from all possible pairwise combinations of individuals was found to be 0·203. The standard error for IA under the assumption of no linkage disequilibrium was calculated to be 0·0008; therefore we are confident that the observed IA is significantly different from zero, which implies that there is some degree of linkage disequilibrium among the loci studied. The IA values for the individual population samples were not significantly different from zero at 4 out of the 20 sample stations, but at 3 of these 4 stations (L2, L19 and AH8) gene diversity was exceptionally low (in the bottom four of the 20 stations studied). As allele frequencies tend towards fixation, the ability of IA to discriminate clonality and panmixis is much diminished, and thus these results must be interpreted with caution. Most importantly, the fact that the majority of IA values from individual stations are non-zero shows that the linkage disequilibrium detected in the whole dataset is not the result of pooling genetically discrete subpopulations, each of which is in linkage equilibrium.

Modified Mantel tests for linkage disequilibrium within 20 randomly selected subsets of 100 individuals of Nodularia were performed. Analysis showed significant linkage disequilibrium between some loci combinations in 17 out of 20 subsets of data (Table 6). Having detected linkage disequilibrium in the data, we compared the observed number of filaments with each possible pairwise combination of alleles at any two loci to the number expected for a population in linkage equilibrium. The combination gvpA-IGS II+rDNA-ITS II was found to occur less often (29 individuals) than expected by chance (69 individuals): this difference is highly significant ({chi}2=25, P=<0·01).


View this table:
[in this window]
[in a new window]
 
Table 6. Summary of randomized reassortment of Nodularia genotypes

 
Gene diversity, a convenient measure of the probability that any two filaments picked at random from the population will have different alleles at a given locus, was calculated for the population as a whole. Values were found to range from 0·06 for gvpA-IGS to 0·21 for PC-IGS. The total gene diversity value summed for all three loci was 0·41, giving a mean value across all three loci of 0·14.


   DISCUSSION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Studies of bacteria, including fluorescent pseudomonads (Haubold & Rainey, 1996 ) and members of the Enterobacteriaceae (Gordon & FitzGibbon, 1999 ), have demonstrated a link between population genetic structure and properties of the local environment or host respectively. In contrast, the spatial distribution of Nodularia genotypes could not be explained by any of the environmental variables examined in this study, including concentrations of key nutrients, salinity, temperature and depth of the water column. The lack of correlation between the physical/chemical conditions at each sampling location and the relative abundance of particular genotypes may be explained in two ways. First, there may not be sufficient variation in conditions across the study region to differentially select between Nodularia genotypes; and second, gene exchange may well be common enough to randomize any association between the non-coding loci studied, and those actually under selection in the environment. The difference between our findings and those of Haubold & Rainey (1996 ) and Gordon & FitzGibbon (1999 ) may reflect the fact that these earlier studies were based on samples collected from discrete habitats, individual leaves (Haubold & Rainey, 1996 ) or mammalian hosts (Gordon & FitzGibbon, 1999 ), whereas we collected samples from a large but contiguous area of the Baltic Sea.

The finding that the most distantly spaced sampling stations were genetically similar (Fig. 4) was initially surprising. A simple explanation for this observation seems to be that for the greatest distances, one is comparing stations at the extreme southwest and northeast of the sampling area. The northern limit of the population was not explored in this study, but our records show that Nodularia was not found at Litorina station 1 (54·53° N, 10·51° E: Fig. 3), just 44 km west of station 2 (Table 1). This indicates that station 2, the station with the lowest gene diversity, may have been close to the southwestern limit of the Nodularia population at the time of sampling. It is possible that the extremes of Nodularia’s range are tolerable only to a subset of genotypes, resulting in lower genetic diversity in these areas. The significant negative correlation between the level of gene diversity at each station and its distance from the centre of the study (Fig. 5) provides some limited support for this hypothesis.

The most significant finding of this study is that genetic exchange almost certainly occurs between Nodularia filaments in the natural population. This is revealed by the fact that all 12 possible allele combinations (genotypes) were found to exist in our sample of the population. An alternative explanation for the occurrence of the 12 observed genotypes is that parallel mutation along each of several purely asexual lineages has by chance generated the same alternative alleles (but no others) many times. Suppose, for example, that the ancestor of all Nodularia in the Baltic Sea today first diverged into two clonal lineages, each of which accumulated its own unique mutations in the PC-IGS region. We must then assume that by chance, the same suite of point mutations (and no others) occurred in the rDNA-ITS of these two independent lineages, and that finally a parallel insertion/deletion event in each of the six resulting lineages gave rise to the two gvpA-IGS alleles, and thus to the 12 genotypes (see Barker et al., 1999 , for an exact description of the differences between the alternative alleles, but note that including one unpublished PC-IGS sequence, we have at least three sequences for each allele type at each locus used in this study). Given the staggering odds against the sequence of events described above actually occurring, genetic exchange clearly provides the most parsimonious explanation of the data. We have also considered the possibility that either gene conversion or gene loss could produce the observed distribution of genotypes. The gvpA and rDNA coding regions are present in multiple copies in at least some Nodularia isolates (all studied so far in the case of gvpA), and gene conversion could therefore explain changes in allele patterns without the need for gene exchange. If we take the rDNA locus first, there are three allele types: type I amplifiable with primer 9408R, type II amplifiable with 9427R and type III amplifiable with both primers. Although it might seem as though a gene conversion event could transform a type III isolate into a type I or II, sequence analysis of type III isolates has shown that the rDNA copies present are not identical to those found in either type I or type II isolates (Barker et al., 1999 ). Gene conversion can not therefore account for the interconversion of any known rDNA types. Similarly, gene conversion is unlikely to explain differences in the gvpA-IGS genotype of our samples: although there are two copies of gvpA, there is just a single copy of the intergenic spacer that separates them. Of course we can not rule out the possibility that there was an ancestral Nodularia which contained all possible alleles at each locus, and that differential loss by gene conversion or deletion gave rise to the 12 types found today, but again gene exchange offers the most parsimonious explanation of the data.

This is the first population-level study suggestive of genetic exchange in a natural cyanobacterial population. The index of association is significantly different from zero, which means that whilst gene exchange could occur in Nodularia, the population is not panmictic. This conclusion is also supported by the randomization tests, which showed that whilst some combinations of filaments showed significantly more linkage disequilibrium than expected by chance, other combinations could not be distinguished from simulated random mating data (Table 6). Unfortunately it is not possible to estimate actual levels of recombination by analysing linkage disequilibrium (Maynard Smith et al., 1993 ; Cohan, 1994 ), as this would require other factors such as the mutation rate to be quantified. What is clear from the data is that recombination is not commonplace enough to completely randomize alleles in the population; if that were the case the index of association would not differ significantly from zero, and the randomization tests would show no significant combinations. There are three possible explanations for the non-zero index of association between loci. First, the loci studied may be closely linked (i.e. lie in close physical proximity); second, certain allelic combinations may be selectively advantageous; and third, the population may comprise a set of clones which have undergone many divisions without any genetic exchange with other genotypes.

Whilst analysis of observed and expected numbers for each pairwise combination of loci showed that rDNA-ITS II+gvpA-IGS II occurred less frequently than expected, we have no reason to suspect that this particular allele combination should have a low fitness value, especially as in the case of the rDNA region, sequence variation has only been detected in the ITS, and not in coding regions (Barker et al., 1999 ). It is possible, however, that selection could be acting on regions tightly linked to those studied. It is unlikely that the three loci chosen are closely linked, but since we do not have any information about their physical location in the Nodularia genome we can not rule out this possibility. The most likely explanation for the observed degree of linkage disequilibrium seems to be that Nodularia undergoes periods of purely asexual growth, during which time certain genotypes grow better than others, but that gene exchange, by unknown mechanisms, occurs frequently enough to ensure that no particular allele at one locus ever occurs exclusively in combination with another from a separate genomic location: such a population structure is described as epidemic by Maynard Smith et al. (1993 ). There have been no other population genetic studies on cyanobacteria with which to compare the results obtained for Nodularia; therefore we have no idea whether the population structure observed for this taxon is typical of filamentous cyanobacteria or of the cyanobacteria as a whole. Given the variation in IA in the genus Bacillus (Istock et al., 1992 ) it seems unwise to extrapolate widely from our findings. Other studies have addressed genetic diversity in aquatic micro-organisms; for example Wise et al. (1996 ) used data from seven polymorphic allozyme loci to show that the mean gene diversity for a natural lotic population of Burkholderia cepacia was 0·52. Unfortunately a useful comparison of genetic diversity based on allozyme and DNA markers respectively is impossible. We hope that our data can form a baseline with which to compare future DNA-marker studies of other aquatic microbes.

That genetic exchange can occur between bacteria in the marine environment is not a novel finding; intra-generic conjugative transfer and transformation at frequencies of around 10-2 and 10-7 respectively have been observed in many in vitro studies with marine water and sediments (see review by Ashelford et al., 1997 ) but all of the studies to date have involved pseudomonads, Escherichia coli or Vibrio spp.

Comparison with data from a research cruise in 1996 (Hayes & Barker, 1997 ) shows that the frequency of the two PC-IGS alleles in the Baltic Sea changed dramatically over 2 years. In 1996, PC-IGS types I/II and III were present in almost equal numbers (n=91 and 88 respectively), but in 1998 type I/II was rare in comparison to type III (n=355 and 2335 respectively). Undoubtedly the abundance of filaments of genotype 9 (Table 1) is the major contributing factor in accounting for this shift in the allelic composition of the Nodularia population. We will not be able to explain such temporal shifts in population structure until we know more about the relative fitness of different Nodularia genotypes under various environmental conditions, and what impact cyanophage and other pathogens may have on dominant genotypes.

In terms of the taxonomy of Nodularia in the Baltic Sea, this study, as with our previous study (Barker et al., 1999 ) indicates that the Baltic has only a single species, if one chooses to apply a biological species concept.

The data presented here lend support to the hypothesis of Rudi et al. (1998 ) that the occurrence of genetic exchange best explains the incongruent pattern of inheritance for the small-subunit rRNA and rbcLX regions of many cyanobacteria. What our study shows is that gene exchange occurs in some cyanobacteria at a frequency great enough not just to show up over lengthy evolutionary time scales as demonstrated already, but also to shape the genetic structure of extant populations. We are now studying temporal changes in allele frequencies in greater detail, and starting to explore the mechanisms which facilitate genetic exchange in Nodularia.


   ACKNOWLEDGEMENTS
 
We thank Annette Richer, Anja Engel, Klaus von Bröckel and the captains and crews of the FS Litorina and the RV Alexander von Humboldt for their assistance. This work was supported by the EC Environment programme (ENV4-CT97-0571), and the Natural Environment Research Council of the United Kingdom (GR9/3851, GR3/12415). J.R.S. is supported by the Wellcome Trust. This is ELOISE publication number 164.


   REFERENCES
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Ashelford, K. E., Fry, J. C., Day, M. J., Hill, K. E., Learner, M. A., Marchesi, J. R., Perkins, C. D. & Weightman, A. J. (1997). Using microcosms to study gene transfer in aquatic habitats. FEMS Microbiol Ecol 23, 81-94.

Barker, G. L. A., Hayes, P. K., O’Mahony, S. L., Vacharapiyasophon, P. & Walsby, A. E. (1999). A molecular and phenotypic analysis of Nodularia (cyanobacteria) from the Baltic Sea. J Phycol 35, 931-937.

Barten, R. & Lill, H. (1995). DNA-uptake in the naturally competent cyanobacterium, Synechocystis sp. PCC-6803. FEMS Microbiol Lett 129, 83-88.

Bolch, C. J. S., Blackburn, S. I., Neilan, B. A. & Grewe, P. M. (1996). Genetic characterisation of strains of cyanobacteria using PCR-RFLP of the cpcBA intergenic spacer and flanking coding regions. J Phycol 32, 445-451.

Bolch, C. J. S., Orr, P. T., Jones, G. J. & Blackburn, S. I. (1999). Genetic, morphological and toxicological variation among globally distributed strains of Nodularia (cyanobacteria). J Phycol 35, 339-355.

Chiang, G. G., Schaefer, M. R. & Grossman, A. R. (1992). Transformation of the filamentous cyanobacterium Fremyella diplosiphon by conjugation or electroporation. Plant Physiol Biochem 30, 315-325.

Cohan, F. M. (1994). Genetic exchange and evolutionary divergence in prokaryotes. Trends Ecol Evol 9, 175-180.

Eardly, B. D., Materou, L. A., Smith, N. H., Johnson, D. A., Rumbaugh, M. D. & Selander, R. K. (1990). Genetic structure of natural populations of the nitrogen-fixing bacterium Rhizobium meliloti. Appl Environ Microbiol 56, 187-194.[Medline]

Feizabadi, M. M., Robertson, I. D., Cousins, D. V., Dawson, D. J. & Hampson, D. J. (1997). Use of multilocus enzyme electrophoresis to examine genetic relationships amongst isolates of Mycobacterium intracellulare and related species. Microbiology 143, 1461-1469.[Abstract]

Gordon, D. M. & FitzGibbon, F. (1999). The distribution of enteric bacteria from Australian mammals: host and geographical effects. Microbiology 145, 2663-2671.[Abstract/Free Full Text]

Hagen, M. J. & Hamrick, J. L. (1996). Population level processes in Rhizobium leguminosarum bv trifolii: the role of founder effects. Mol Ecol 5, 707-714.

Haubold, B. & Rainey, P. B. (1996). Genetic and ecotypic structure of a fluorescent Pseudomonas population. Mol Ecol 5, 747-761.

Hayes, P. K. & Barker, G. L. A. (1997). Genetic diversity within Baltic Sea populations of Nodularia (cyanobacteria). J Phycol 33, 919-923.

Istock, C. A., Duncan, K. E., Ferguson, N. & Zhou, X. (1992). Sexuality in a natural population of bacteria – Bacillus subtilis challenges the clonal paradigm. Mol Ecol 1, 95-103.[Medline]

Komárek, J., Hübel, M., Hübel, H. & Smarda, J. (1993). The Nodularia studies. 2. Taxonomy. Arch Hydrobiol Algol Stud 68, 1-25.

Lorenz, M. G. & Wackernagel, W. (1994). Bacterial gene-transfer by natural genetic-transformation in the environment. Microbiol Rev 58, 563-602.[Abstract]

Mantel, N. (1967). The detection of disease clustering and a generalised regression approach. Cancer Res 27, 209-220.[Medline]

Maynard Smith, J. (1995). Do bacteria have population genetics? In Population Genetics of Bacteria (Society for General Microbiology Symposium no. 52), pp. 1-12. Edited by S. Baumberg, J. P. W. Young, E. M. H. Wellington & J. R. Saunders. Cambridge: Cambridge University Press.

Maynard Smith, J., Smith, N. H., O’Rourke, M. & Spratt, B. G. (1993). How clonal are bacteria? Proc Natl Acad Sci USA 90, 4384-4388.[Abstract]

Muropastor, A. M., Kuritz, T., Flores, E., Herrero, A. & Wolk, C. P. (1994). Transfer of a genetic-marker from a megaplasmid of Anabaena sp. strain PCC-7120 to a megaplasmid of a different Anabaena strain. J Bacteriol 176, 1093-1098.[Abstract]

Nei, M. (1972). Genetic distance between populations. Am Nat 106, 283-292.

Nei, M. (1987). Estimation of average heterozygosity and genetic distance from small numbers of individuals. Genetics 89, 583-590.

Neilan, B. A., Jacobs, D. & Goodman, A. E. (1995). Genetic diversity and phylogeny of toxic cyanobacteria determined by DNA polymorphism within the phycocyanin locus. Appl Environ Microbiol 61, 3875-3883.[Abstract]

Rudi, K., Skulberg, O. M. & Jakobsen, K. S. (1998). Evolution of cyanobacteria by exchange of genetic material among phyletically related strains. J Bacteriol 180, 3453-3461.[Abstract/Free Full Text]

Silva, C., Eguiarte, L. E. & Souza, V. (1999). Reticulated and epidemic population genetic structure of Rhizobium etli biovar phaseoli in a traditionally managed locality in Mexico. Mol Ecol 8, 277-287.

Singh, D. T., Bagchi, S. N., Modi, D. R. & Singh, H. N. (1987). Evidence for intergeneric transformation in filamentous, diazotrophic cyanobacteria. New Phytol 107, 347-356.

Sode, K., Tatara, M., Takeyama, H., Burgess, J. G. & Matsunaga, T. (1992). Conjugative gene-transfer in marine cyanobacteria Synechococcus sp., Synechocystis sp. and Pseudanabaena sp. Appl Microbiol Biotechnol 37, 369-373.[Medline]

Spratt, B. G., Smith, N. H., Zhou, J., O’Rourke, M. & Feil, E. (1995). The population genetics of the pathogenic Neisseria. In Population Genetics of Bacteria (Society for General Microbiology Symposium no. 52), pp. 143-160. Edited by S. Baumberg, J. P. W. Young, E. M. H. Wellington & J. R. Saunders. Cambridge: Cambridge University Press.

Stevens, J. R. & Tibayrenc, M. (1995). Detection of linkage disequilibrium in Trypanosoma brucei isolated from tsetse flies and characterised by RAPD analysis and isoenzymes. Parasitology 110, 181-186.[Medline]

Stevens, J. R. & Tibayrenc, M. (1996). Trypanosoma brucei s.l.: evolution, linkage and the clonality debate. Parasitology 112, 481-488.[Medline]

Tandeau de Marsac, N., Mazel, D., Bryant, D. A. & Houmard, J. (1985). Molecular-cloning and nucleotide-sequence of a developmentally regulated gene from the cyanobacterium Calothrix-PCC-7601 – a gas vesicle protein gene. Nucleic Acids Res 13, 7223-7236.[Abstract]

Thiel, T. & Wolk, C. P. (1987). Conjugal transfer of plasmids to cyanobacteria. Methods Enzymol 153, 232-243.[Medline]

Trott, D. J., Oxberry, S. L. & Hampson, D. J. (1997). Evidence for Serpulina hyodysenteriae being recombinant with an epidemic population structure. Microbiology 143, 3357-3365.[Abstract]

Walsby, A. E., Hayes, P. K. & Boje, R. (1995). The gas vesicles, buoyancy and vertical distribution of cyanobacteria in the Baltic Sea. Eur J Phycol 30, 87-94.

Whittam, T. S. (1995). Genetic population structure and pathogenicity in enteric bacteria. In Population Genetics of Bacteria (Society for General Microbiology Symposium no. 52), pp. 217-245. Edited by S. Baumberg, J. P. W. Young, E. M. H. Wellington & J. R. Saunders. Cambridge: Cambridge University Press.

Wilmotte, A. (1994). Molecular evolution and taxonomy of the cyanobacteria. In The Molecular Biology of Cyanobacteria, pp. 1-25. Edited by D. A. Bryant. Dortrecht: Kluwer.

Wilson, W. H. & Mann, N. H. (1997). Lysogenic and lytic viral production in marine microbial communities. Aquat Microb Ecol 13, 95-100.

Wise, M. G., McArthur, J. V., Wheat, C. & Shimkets, L. J. (1996). Temporal variation in genetic diversity and structure of a lotic population of Burkholderia (Pseudomonas) cepacia. Appl Environ Microbiol 62, 1558-1562.[Abstract]

Received 12 April 2000; revised 21 July 2000; accepted 31 July 2000.