Experimental evolution of the megaplasmid pMPPla107 in Pseudomonas stutzeri enables identification of genes contributing to sensitivity to an inhibitory agent

Abstract

Horizontally transferred elements, such as plasmids, can burden host cells with various metabolic and fitness costs and may lead to other potentially detrimental phenotypic effects. Acquisition of the Pseudomonas syringae megaplasmid pMPPla107 by various Pseudomonads causes sensitivity to a growth-inhibiting substance that is produced in cultures by Pseudomonads during growth under standard laboratory conditions. After approximately 500 generations of laboratory passage of Pseudomonas stutzeri populations containing pMPPla107, strains from two out of six independent passage lines displayed resistance to this inhibitory agent. Resistance was transferable and is, therefore, associated with mutations occurring on pMPPla107. Resequencing experiments demonstrated that resistance is likely due to a large deletion on the megaplasmid in one line, and to a nonsynonymous change in an uncharacterized megaplasmid locus in the other strain. We further used allele exchange experiments to confirm that resistance is due to this single amino acid change in a previously uncharacterized megaplasmid protein, which we name SkaA. These results provide further evidence that costs and phenotypic changes associated with horizontal gene transfer can be compensated through single mutational events and emphasize the power of experimental evolution and resequencing to better understand the genetic basis of evolved phenotypes.

This article is part of the theme issue ‘The secret lives of microbial mobile genetic elements’.

1. Introduction

Plasmids are extrachromosomal replicons that can often move between bacterial cells through conjugation, and that are a major contributor to horizontal gene transfer (HGT) events in bacteria. Thousands of genes can be exchanged via HGT in a single transfer and this potentially opens up new niches for each organism through the acquisition of genes encoding proteins involved in metabolism, antibiotic resistance, virulence and symbiosis [1–6]. Although plasmids can provide a variety of advantages to a bacterial cell in a given environment, HGT can also engender general fitness costs and can lead to additional phenotypic changes that could be costly under specific environmental contexts [7,8]. The existence of such costs has led some to generally question how plasmids are maintained within populations over evolutionary time given plasmid–chromosome conflicts, a discussion that is often referred to as the ‘plasmid paradox’ [9,10]. Identification of compensatory mutations selected to ameliorate plasmid–chromosome conflicts during laboratory passage can significantly inform our understanding of the mechanistic basis of these conflicts and may, therefore, ultimately provide deeper molecular insights into the costs of HGT as well as the ‘plasmid paradox’ [7,8].

We have previously shown that acquisition of the Pseudomonas syringae megaplasmid pMPPla107 by Pseudomonas stutzeri sensitizes this strain background to the presence of an inhibitory agent that has bacteriostatic properties [11–13]. Sensitivity is found in pMPPla107's original host strain P. syringae pv. lachrymans 107 and can be transferred to various Pseudomonas spp. upon their acquisition of pMPPla107, thus indicating that the phenotype is linked to the acquisition of pMPPla107 [12]. Production of this inhibitory agent is conserved across measured Pseudomonas spp., appears correlated with Pseudomonas physiology, and is likely linked to an essential process as all viable strains within a P. aeruginosa transposon library maintained activity [13].

Although the acquisition of plasmids by new host backgrounds often creates metabolic, physiological and fitness costs, previous research has shown that various types of compensatory mutations occur rapidly on either the chromosome or plasmid, with amelioration of costs enabling the persistence of plasmids [14–17]. For instance, experiments on P. fluorescens with the mercury-resistant pQBR103 demonstrate that compensatory mutations in gacA/gacS occur in strains with and without selection using mercury [14]. Moreover, mutations in two helicases and an RNA polymerase subunit resulted in host dependence on the plasmid RP4 while also increasing the uptake of additional plasmids [15]. Therefore, compensatory mutations may not only explain plasmid persistence mechanisms but also increased plasmid promiscuity.

For these reasons and to better understand the genetic basis of previously described phenotypic costs associated with pMPPla107 [11], we carried out the experimental passage of P. stutzeri populations under conditions that select for maintenance of pMPPla107. Our goal with these passage experiments was to identify strain backgrounds that evolve resistance to this inhibitory agent with the hope that identification of compensatory mutations would provide better understanding of the genetic basis of these costs.

2. Methods

(a) . Laboratory passage experiment

A frozen stock of strain DBL408 [11] was streaked to salt water lysogeny broth (SWLB) agar supplemented with rifampicin (50 ng µl⁻¹) and tetracycline (10 ng µl⁻¹). Hereafter, every instance mentioned for SWLB liquid and solid media will also include supplementation with these antibiotics at these concentrations unless otherwise noted. From this initial culture plate, six independent single colonies of P. stutzeri strain DBL408 were picked into 2 ml cultures of SWLB in 5 ml polypropylene tubes with caps. These cultures were grown under shaking (220 r.p.m.) conditions at 27°C for 2 days, at which point a subset of these cultures was frozen in 40% glycerol at −80°C and labelled as ‘passage 0’ while a 1 : 1000 (cells : media) dilution was carried out into fresh 2 ml of SWLB supplemented with the same antibiotics as above. Each passage, cells were plated to SWLB agar plates to observe colony morphology in case of contamination. Tetracycline in the media selects for maintenance of the megaplasmid in strain DBL408, while rifampicin protects against contamination from additional sources. Although specific experiments were not performed to evaluate this point directly, it is likely that megaplasmids would be lost from these strains and potentially whole populations during early passages given the cost of megaplasmid carriage in P. stutzeri [11]. Every 10 passages, a 750 µl sample of the culture was mixed to 40% final concentration with glycerol and stored at −80°C. This passage process was repeated for approximately 500 generations of growth (log₂ of 1000 = 9.96 divisions per passage; 50 passages total).

After 50 passages, a sample from each independent evolutionary population was streaked onto SWLB agar plates. Two single colonies (A-500 and B-500) were picked from each evolutionary passage line and frozen as single strain isolates from each passage population. We focused on isolating single colonies, rather than on whole population sequencing, with the goal of linking identifying distinct phenotypic changes to underlying genotypic changes to better understand the underlying molecular mechanisms rather than on profiling evolutionary dynamics across whole populations. In this way, our experiments described herein treat laboratory passage like a phenotypic screen or selection in a microbial genetics experiment [18].

Similar procedures to the above were used to isolate single colony strains from whole populations at generations 0 and 100 from population 4, except that frozen stock cultures were streaked to SWLB agar plates and approximately 10 colonies were picked in each instance as a rough way to sample population frequencies.

(b) . Inhibitory agent sensitivity test

We isolated single colonies from whole populations of six evolved lines as a means to facilitate linking observed phenotypic changes in these lines to underlying genetic events. We followed previously described protocols [11,13] to test the activity of the inhibitory agent against isolates from each of the six evolved populations. Briefly, overlays were prepared by diluting an overnight culture into fresh media and incubating this new culture on a shaker at 27°C for 4 h with one exception. In the case of skaA allelic exchange overlays, we report the results after 5 h of growth in this manuscript but also include data from 4 h incubation in electronic supplementary material, files S6–S10. Cells from this 4-h culture were then mixed with 0.4% molten agar and plated on King's media B (KB) agar media. This 0.4% agar overlay was allowed to solidify for approximately 15 min, at which point supernatant was added. The inhibitory agent was collected by growing P. stutzeri for 24–48 h, centrifuging cells at 10 000×g for 5 min and sterilizing supernatants through a 0.22 µm filter. After sterilization, 10 µl of supernatants were spotted onto the overlay plate and allowed to dry. Overlay plates were grown at 27°C for approximately 24 h, at which point zones of inhibition were observed. Raw photos from overlays for most experiments are included as electronic supplementary material, figures S1–S10, with each assay taking place on 30 mm petri dishes for scale.

(c) . Conjugation of evolved megaplasmids into ‘ancestral’ P. stutzeri

An ‘ancestral’ strain of P. stutzeri (DBL388), which has undergone fewer than 5 laboratory passages, was mixed with strains 3B-500, 4B-500 and 5B-500 for megaplasmid transfer experiments. DBL388 was created through Tn7 transposition of a transposon from plasmid AKN34 [19] into DBL332 [12] and is resistant to gentamicin. For conjugation experiments, overnight liquid cultures of each strain were mixed 1 : 1 and then centrifuged at 3000×g for 3 min. The supernatant was removed without disturbing the pellet, and pellets were washed and resuspended in 1 ml of 10 mM MgCl₂. Strain mixtures were centrifuged and washed one additional time, after which 100 µl of resuspended cells were spread on KB plates with rifampicin (50 ng µl⁻¹) and incubated for at least 48 h at 27°C. After 48 h, strain mixtures were resuspended in 10 mM MgCl₂, and plated on KB agar plates supplemented with gentamicin (10 ng µl⁻¹) and tetracycline (10 ng µl⁻¹). Gentamicin- and tetracycline-resistant colonies arising on this plate were picked to liquid culture and underwent diagnostic PCR for the presence of pMPPla107 using primers from [20].

(d) . Genome sequencing and annotation

For sequencing experiments, a single colony of each strain was picked to 2 ml SWLB media and grown overnight, shaking at 27°C. DNA samples were extracted from these cultures using a Promega Wizard kit (Madison, WI) with RNAse treatments. P. stutzeri strain DBL408 (the progenitor strain of all experimental lines) as well as strains 1B-500, 2B-500, 4B-500 and 6B-500 were sequenced using 100 bp paired end reads on an Illumina HiSeq by the University of Arizona sequencing core (SRX11065275, SRX5391491, SRX5391490, SRX5391492, SRX5391486, respectively). Pseudomonas stutzeri lines 3B-500 and 5B-500 were sequenced using 250 bp paired end reads on an Illumina MiSeq by MicrobesNG (Birmingham, UK; SRX5391493, SRX5391487, respectively). For the identification of polymorphisms, we used gene annotations of pMPPla107 from a previous publication ([1], Accession: NZ_CP031226.1) and the annotations from the P. stutzeri 28a24 reference sequence ([21], Accession: CP007441.1). We additionally isolated a single colony from the passage 0 stock for experimental population 4, referred to herein as 4A-0, and the exconjugant strain DBL619 (arising from a mating between 4B-500 and DBL388). Genomic DNA for these isolates were isolated as above and sequenced using 150 bp paired end reads on an Illumina MiSeq by SNPsaurus (Eugene, OR; SRX11065277, SRX11065276, respectively). Lastly, we isolated single colonies from P. stutzeri strains DBL494 and DBL1802, and extracted genomic DNA as above. Genomic DNA for these two strains was sequenced using 150 bp paired end reads on an Illumina NovaSeq by MiGS (Pittsburgh, PA; SRX11559214 and SRX11559215, respectively).

We further assembled reads de novo from strains 4A-0, 4B-500 and 5B-500 using unicycler v. 0.8.4 and with default parameters. Contigs for these assemblies can be found in electronic supplementary material, file S2, 4A-0; file S3, 4B-500; file S4, 5B-500. We note that our best estimate based on read depth and coverage is that megaplasmid pMPPla107 is found in approximately the same copy number as the chromosome and there is no indication based on these new assemblies that the copy number changed over the course of the experimental passage.

(e) . Mapping reads and calling variants

Illumina reads from all six evolved lines were mapped to the pMPPla107 reference sequence (accession CP031226.1) using Breseq v. 0.35.5 [22] with default parameters. We only considered variants that were strongly supported for analyses within this manuscript.

(f) . Allelic replacement of skaA

To definitively demonstrate that the skaA allele from P. stutzeri 5B-500 is causative of the observed change in sensitivity to supernatants, we moved the P. stutzeri 5B-500 allele into a naive strain background. Briefly, a dsDNA block containing the P. stutzeri 5B-500 allele as well as a silent change in skaA that introduces an Eco53kI restriction site was obtained from IDT (Coralville, IA) and is found as file S1 in the electronic supplementary material. BP clonase (Invitrogen; Waltham, MA) was used to recombine this dsDNA block into the pDONR207 entry vector to create plasmid pDBL100. Then, LR clonase (Invitrogen; Waltham, MA) was used to recombine this dsDNA block from plasmid pDBL100 into destination vector pMTN1907 to create plasmid pDBL101. Plasmid pDBL101 was transformed into Escherichia coli strain S17 (to create strain DBL1804), and conjugated into P. stutzeri strain DBL494. Strain DBL494 is a nalidixic acid- and gentamicin-resistant derivative of P. stutzeri strain DBL492, which is a nalidixic acid (100 ng ul⁻¹)-resistant derivative of strain 28a24. To create strain DBL494, a mariner transposon providing gentamicin resistance was transposed into the pMPPla107 megaplasmid in strain DBL453 and then conjugated into strain DBL492. The transposon in this strain has been localized to the megaplasmid by sequencing [23].

A tetracycline-resistant strain was isolated from the conjugation between E. coli DBL1804 and DBL494, in which plasmid pDBL101 was integrated into the megaplasmid in DBL494. A single tetracycline-resistant colony was picked to 2 ml plain KB media and grown overnight, at which point dilutions were plated onto KB media agar supplemented with 10% sucrose. Sucrose-resistant colonies were screened for tetracycline resistance and remaining tetracycline sensitive colonies were screened by PCR using forward primer 5′0-GCCTGGGTGACACCTATCAG-3′ and reverse primer 5′-AATGCTCAGCTGCAGTCGAT-3′ using an annealing temperature of 55°C and an extension time of 1 : 30. Colonies were considered to contain the allelic replacement if the PCR product from this reaction was not cut by Eco53kI.

(g) . Synteny plots

We used SynMap2 with the LAST algorithm and default parameters to compare the sequences of ancestral pMPPla107 and pMPPla107-4B500 [24]. DAGChainer Options were: nucleotide distance, −D = 20, and −A = 5. The Tandem duplication distance was set to 10 and the C-score was set to 0.

(h) . Gene function and structure predictions

To predict functional characteristics, we input the amino acid sequence of SkaA as input into a PsiBlast search of the NCBI non-redundant protein sequence database (date of search last search: 19 July 2021) [25]. PsiBlast was run for four iterations, at which point no new protein matches were added as results. The amino acid sequence of SkaA was also used as an input to the Phyre2 web server [26], with selection of the ‘intensive’ setting. Lastly, we used a Google collaboration implemented version of AlphaFold 2 to predict three-dimensional structure of the protein [27,28]. Briefly, sequences are picked through BLAST and aligned using MMseqs2. The resulting .pdb file can be found as electronic supplementary material, file S5, and was visualized through the Protein Data Bank and Mol* [29,30].

(i) . Supplemental data

All supplemental data for this manuscript can be found on Figshare at doi:10.6084/m9.figshare.15072537. Electronic supplementary material, table S1 lists all strains and plasmids used in this manuscript and includes citations for [31,32]. Electronic supplementary material, figures S1–S10 are raw pictures used for overlay experiments. Electronic supplementary material, figure S11 is a Mauve alignment of the 4A-0, 4B-500 and pMPPla107 megaplasmid sequences. Electronic supplementary material, file S1 is a .fasta file for the sequence that was ordered from IDT and used to create vectors for allelic exchange experiments with skaA. Electronic supplementary material, files S2–S4 are .fasta files for genome assemblies of strains 4A-0, 4B-500 and 5B-500. Electronic supplementary material, file S5 is a .pdb file for the SkaA structure prediction.

3. Results

(a) . Strains from 2 of 6 laboratory passage populations gain resistance to a previously described inhibitory agent

We previously identified that acquisition of plasmid pMPPla107 by P. stutzeri sensitizes this strain to an unknown inhibitory agent produced by numerous Pseudomonas strains [11]. To identify strain backgrounds that were resistant to this agent, we screened for the presence of inhibition in single colony isolates from these evolved populations. Single colony isolates from two out of six lines (referred to from here on as 4B-500 and 5B-500) revert to the non-pMPPla107 phenotype and demonstrate resistance to this inhibitory agent (figure 1).

Figure 1.

Testing inhibitory agent on all six evolved lines reveals two resistant lines. Six populations carrying pMPPla107 were passaged under laboratory conditions for 500 generations. Single colonies isolated from these populations were tested for sensitivity against inhibitory agent found in Pseudomonas spp. supernatants. Isolates from lines 4B-500 and 5B-500 revert to a non-pMPPla107 phenotype where a zone of inhibition is not present, indicating resistance to the inhibitory agent. All overlays were plated after 4 h of growth in KB and spotted with 10 µl of P. stutzeri filter-sterilized supernatants. The number (1, 2, 3, …) indicates the individual lines and B indicates the second of two isolates taken at Generation 500. All images are representative of three biological replicates.

(b) . Mutations providing resistance to the inhibitory agent are transferable and localized to the megaplasmid

In previous evolutionary studies focusing on plasmids, the burden of plasmid acquisition selected for compensatory mutations to arise on both the host chromosome [9,14,15,33] and on the plasmids themselves [17]. To identify whether resistance mutations to this inhibitory agent occurred on the chromosome or megaplasmid, we carried out conjugation experiments to move megaplasmids from 3B-500 (sensitive), 4B-500 (resistant) and 5B-500 (resistant) to an ancestral P. stutzeri background. Exconjugants with the ‘ancestral’ chromosome and evolved megplasmids from 4B-500 and 5B-500 (but not 3B-500) demonstrate resistance to this inhibitory agent (figure 2). Therefore, mutations providing resistance to inhibition from lines 4B-500 and 5B-500 can be positionally localized to the evolved megaplasmids.

Figure 2.

Exconjugants of pMPPla107 from populations 4 and 5 into an ancestral chromosomal background transfers resistance to the inhibitory agent. Given that strains 4B-500 and 5B-500 were known to display resistance against the inhibitory agent, we conjugated megaplasmids from these strains into an unevolved P. stutzeri background DBL388. We additionally used conjugations with the megaplasmid from 3B-500 as a sensitive (positive) control as we knew this evolved strain was still sensitive. While a megaplasmid from 3B-500 maintains sensitivity to the inhibitory agent in the DBL388 background, megaplasmids from 4B-400 and 5B-500 transfer resistance. All overlays were plated after 4 h of growth in KB and spotted with 10 µl of P. stutzeri filter-sterilized supernatants. All images are representative of three biological replicates. Raw image files can be found in electronic supplemental material at Figshare (doi:10.6084/m9.figshare.15072537.v3): DBL618 is electronic supplementary material, figure S1, DBL619 is electronic supplementary material, figure S2, DBL620 is electronic supplementary material, figure S3. (Online version in colour.)

(c) . Independent mutations likely provide resistance to the unknown inhibitory agent

To identify the resistance mutations in lines 4B-500 and 5B-500, we resequenced and analysed genomes from single colony isolates arising from each of the six laboratory passage strains of P. stutzeri after 500 generations. We further resequenced the generation 0 isolate from population 4, 4A-0. Although mutations appear to have arisen on megaplasmids in all lines except 1B-500 and 2B-500, lines 4B-500 and 5B-500 carry single mutations that are not present in any of the other strains (table 1). 4B-500 contains a silent mutation in an uncharacterized gene (PLA107_033875), while 5B-500 contains a nonsynonymous mutation in an uncharacterized gene (PLA107_029840). Furthermore, there is a large deletion in 4B-500 that is not present in any other strain (see below).

Table 1.

Mutations occurring on megaplasmid pMPPla107 in experimental passage strains. Amino acid annotation in green is silent and those in blue are missense mutations. Nucleotides that are changed compared to wild type are highlighted in red in the codon.

(d) . A 368 kb deletion in pMPPla107 is strongly linked to supernatant resistance in strain 4B-500

Conjugation of the megplasmid from 4B-500 into a naive background strongly suggests that mutations providing resistance to the inhibitory agent are associated with and colocalized to the megaplasmid (figure 2).

Analysis of the genome from isolate 4B-500 shows that it contains a large deletion of approximately 368 kb, which occurred within pMPPla107 between approximate positions 131–499 kb (figure 3a). This deletion region includes 451 predicted genes, 53 of which have some functional annotation and two tRNA loci (table 2). Interestingly, there appear to be no clearly repetitive or overlapping sites at the ends of the deletion site as visualized by CoGe [24] (https://genomevolution.org/r/uboj).

Figure 3.

A large deletion occurs early during the passage of population 4 and provides resistance to the inhibitory agent. (a) SynMap dotplot visualizes the large deletion occurring from 131 to 499 kb in the evolved 4B pMPPla107 as a large shift across the x-axis. The remaining portions of the sequences maintain perfect synteny, indicating that a clean deletion occurred. The x-axis is ancestral pMPPla107 gene order where x_{1, …, N} = gene_{1, …, N} and the y-axis is the line 4B evolved pMPPla107 gene order where y_{1, …, N} = gene_1…N.]. (b) We tested single colony isolates from frozen stocks of the passage line 4 populations at generations 0 and 100 for sensitivity to the inhibitory agent. We found that resistance to the inhibitory agent is present in the generation 0 isolate from population 4. This resistance is maintained in generation 100 (shown) through generation 500 and is the only unique mutation in pMPPla107 other than a synonymous SNP. All overlays were plated after 4 h of growth n KB and spotted with 10 µl of P. stutzeri filter-sterilized supernatants. All images are representative of three biological replicates. Raw image files can be found in electronic supplementary material at Figshare (doi:10.6084/m9.figshare.15072537.v3): DBL408 is electronic supplementary material, figure S4, 4A-0 is electronic supplementary material, figure S5, 4A-100 is electronic supplementary material, figure S6. (Online version in colour.)

Table 2.

Annotations of genes deleted from the megaplasmid in strain 4B-500.

annotation	start	end	direction
HNH endonuclease	131 589	132 017	reverse
IcmE	135 202	136 674	reverse
IcmK	136 674	137 732	reverse
ribonucleotide-diphosphate reductase subunit beta	142 392	143 558	forward
nucleotidyl transferase AbiEii/AbiGii toxin family protein	144 218	145 108	reverse
endonuclease	149 053	149 961	forward
RtcB family protein	149 987	151 210	forward
transglycosylase SLT domain protein	152 996	153 736	forward
XRE family transcriptional regulator	162 068	162 529	forward
tRNA-Arg	162 927	163 003	forward
HDOD domain-containing protein	166 715	168 022	reverse
DUF805 domain-containing protein	216 222	216 527	reverse
RNA 2′-phosphotransferase	220 194	220 772	reverse
XRE family transcriptional regulator	227 208	227 495	reverse
SMI1/KNR4 family protein	232 488	232 898	reverse
DUF1127 domain-containing protein	269 304	269 519	forward
DUF21 domain-containing protein	269 566	270 537	forward
3-oxoacyl-ACP synthase	273 913	275 001	forward
NAD-dependent epimerase/dehydratase family protein	275 411	276 382	forward
MBL fold metallo-hydrolase	276 375	277 217	forward
tRNA-Leu	282 629	282 713	forward
XRE family transcriptional regulator	284 257	284 511	forward
ASCH domain-containing protein	284 869	285 198	forward
adenylate-forming enzyme	292 503	293 810	forward
Arc family DNA-binding protein	304 934	305 476	forward
RND transporter	312 173	313 522	reverse
ParE	328 197	330 101	forward
ParC	330 900	333 152	forward
DUF721 domain-containing protein	335 955	336 419	reverse
SH3 domain-containing protein	338 158	339 000	forward
DUF4343 domain-containing protein	358 704	359 468	forward
thioredoxin TrxA	368 277	368 594	reverse
DUF541 domain-containing protein	375 046	375 759	forward
HNH endonuclease	378 522	379 406	forward
XRE family transcriptional regulator	403 681	405 477	forward
DUF262 domain-containing protein	415 714	416 247	forward
transposase	433 226	434 356	reverse
regulatory protein RecX	441 216	441 695	forward
FMN-binding glutamate synthase family protein	444 888	446 483	reverse
tryptophan–tRNA ligase	450 046	451 407	reverse
DUF4165 domain-containing protein	452 760	456 818	forward
SPFH domain-containing protein	468 612	469 550	reverse
Ig-like domain-containing protein	473 955	474 710	reverse
multidrug efflux RND transporter permease subunit	475 218	478 319	reverse
efflux RND transporter periplasmic adaptor subunit	478 326	479 429	reverse
DUF3828 domain-containing protein	480 389	480 847	forward
PH domain-containing protein	481 802	482 416	forward
conjugal transfer protein TraO	484 704	485 081	reverse
non-canonical purine NTP pyrophosphatase	487 114	487 686	forward
DUF4031 domain-containing protein	488 336	488 689	reverse
oxidoreductase	48 9082	489 852	forward
OfxX fusion product	493 512	494 291	reverse

To further characterize this mutation, we sequenced a single colony isolate from the passage 0 stock of population line 4 (4A-0). Strain 4A-0 is resistant to this inhibitory agent (figure 3b), and sequencing demonstrates that this line also contains a large deletion on the megaplasmid (electronic supplementary material, file S2 and figure S11). Although these results indicate that the deletion in line 4B is responsible for resistance to the inhibitory agent, the large size of the deletion and the density of genes within this region make it difficult to discern which gene(s) are responsible for the resistance phenotype in this region. Resequencing comparisons of strain 4A-0 to DBL408 do not indicate that any additional mutations (besides the large deletion) have occurred in this strain in the megaplasmid (table 1). Depth and coverage assessments from assemblies of megaplasmids from strains 4A-0 and 4B-500 indicate that copy number remains approximately 1× and it has not dramatically increased or decreased from that in strain DBL408 (electronic supplementary material, files S2 and S3).

(e) . A single polymorphism present on the megaplasmid from strain 5B-500 is responsible for supernatant resistance

Conjugation of the megaplasmid from 5B-500 into a naive background strongly suggests that mutations providing resistance to the inhibitory agent are associated with and colocalized to the megaplasmid (figure 2).

Resequencing of strain 5B-500 demonstrated that a single polymorphism occurs on the megaplasmid from this strain (table 1). Depth and coverage assessments from assemblies of the megaplasmid from strain 5B-500 indicate that copy number remains approximately 1× and it has not dramatically increased or decreased from that in strain DBL408 (electronic supplementary material, file S4).

The mutation identified on the megaplasmid in strain 5B-500 occurs at position 57 137 bp and leads to a nonsynonymous mutation, changing glutamate to a lysine (395 E > K) in a previously uncharacterized protein (PLA107_029840). Data from conjugation experiments and resequencing data strongly suggest that this polymorphism eliminates the sensitivity phenotype seen by strains that have acquired pMPPla107, thus we name this gene skaA for supernatant killing activity A. Given that the megaplasmid polymorphism identified in strain 5B-500 occurs outside the deletion region found in the 4B-500 megaplasmid, this also strongly implies that two separate compensatory strategies exist within pMPPla107 to provide resistance to the unknown inhibitory agent.

To definitively demonstrate that the single mutation present in skaA in P. stutzeri 5B-500 is causative of the change in supernatant sensitivity of this strain, we swapped alleles of skaA from 5B-500 into an otherwise ‘clean’ P. stutzeri genomic background containing the pMPPla107 megaplasmid (DBL494). As one can see in figure 4, the replacement of the wild-type allele in DBL494 with the skaA allele from 5B-500 leads to a loss of sensitivity to supernatant treatment under the measured conditions.

Figure 4.

Allelic replacement of skaAE395 K provides resistance to inhibitory compounds in Pseudomonas supernatants. We recreated the skaAE395 K allele in a naive chromosomal and megaplasmid background (strain DBL494). A zone of inhibition can be seen for strain DBL494 but not for strain DBL1802. Strain DBL494 contains a wild-type allele of skaA on the megaplasmid while the skaAE395 K mutation has been recombined into the megaplasmid in strain DBL1802. Overlays were plated after 5 h of growth in KB liquid and spotted with 20 µl of P. stutzeri filter-sterilized supernatants. All images are representative of three biological replicates for these strains. Raw image files can be found in electronic supplementary material at Figshare (doi:10.6084/m9.figshare.15072537.v3): DBL494 is electronic supplementary material, figure S8, DBL1802 is electronic supplementary material, figure S10.

(f) . Protein predictions for SkaA

SkaA is characterized as a ‘hypothetical protein’ in the current annotation of the pMPPla107 megaplasmid, and does not appear to match any well-characterized proteins found in the current ( July 2021) version of the NR database at NCBI. The only PsiBlast hits that show relatively high sequence similarity to SkaA from plasmid pMPPla107 are found in other Pseudomonas genomes and are likely present on unconfirmed copies of other megaplasmids from the pMPPla107 family. However, SkaA also displays more distant (approx. 18% identity, approx. 37% positives) matches to ‘hypothetical’ proteins from various Vibrio and Staphylococcus species (e.g. WP_094133102.1).

Alphafold2 predicts that SkaA is composed of numerous alpha helices, potentially split up into roughly three domains (figure 5). Results from the Phyre2 server reinforce this prediction as there are no high quality (greater than 50% confidence) matches and all lower quality matches are proteins with domains composed of alpha helices. The aspartate to lysine change in SkaA from P. stutzeri 5B-500 is predicted to occur within one of the numerous predicted alpha helix domains of the protein (figure 5).

Figure 5.

Predicted protein structure of SkaA. We used Alphafold2 to predict the protein structure of SkaA. The position of the E395 K change in the megaplasmid of 5B-500 is denoted with an arrow. (Online version in colour.)

4. Discussion

Acquisition of megaplasmid pMPPla107 by many Pseudomonas strains sensitizes these strains to growth inhibition when treated with supernatants from other Pseudomonas cultures, and we used laboratory passage to isolate and identify mutations associated with resistance to this inhibitory agent. Although strains of P. stutzeri containing pMPPla107 are normally sensitized to the presence of a (currently) unidentified inhibitory agent produced by a variety of Pseudomonas strains under normal growth conditions, isolates from two of six experimental populations evolved resistance to this inhibition over 500 generations of passage. Numerous studies have found that compensatory mutations to plasmid carriage often occur on the chromosome (e.g. [14]), but we found that both mutations providing resistance (in lines 4B-500 and 5B-400/5B-500) likely occur on the megaplasmid.

Sequencing of line 4B-500 demonstrated that this line contains a 368 kb deletion within a region on the megaplasmid that we have previously identified as demonstrating extensive divergence between similar megaplasmids [1]. Since this deletion was present during the initial passage of this population (electronic supplementary material, figure S11), which was initiated from a single colony picked from the strain DBL408 stock, it is possible that this deletion was actually present and circulating as a polymorphism prior to the initiation of all experimental lines. This pattern suggests that this part of the megaplasmid is a potential cargo region where expendable genes may be more likely to provide benefits in certain environments rather than necessary genes for maintenance or transmission. Some of the genes found within this region include efflux pumps, antitoxins and multidrug resistance proteins, all of which could possibly affect resistance to the inhibitory agent (table 2). This region also includes genes annotated as parE and parC loci, which possibly contribute to contrasting sensitivity/resistance phenotypes to quinolone antibiotics by different Pseudomonad strain backgrounds [34]. Lastly, this region also appears to encode an Arc-like DNA-binding protein and two tRNA loci. It is unclear from our current data which of the hundreds of genes in this region is responsible for increased sensitivity to the inhibitory agent. We note that this is another example where large deletions on megaplasmids are shown to contribute to the amelioration of detrimental phenotypic consequences from these plasmids, and highlight that this trend suggests megaplasmid amelioration to new host strains may often occur through large scale deletion of unnecessary genes [35].

We have provided multiple lines of evidence suggesting that a single mutation on the megaplasmid from line 5B-500 imparts resistance to the inhibitory agent. This mutation generates a nonsynonymous change in an uncharacterized open reading frame, which we now formally name skaA. It is still unclear how skaA interacts with inhibitory agent or how the 395 E > K SNP changes these interactions, and protein structure predictions and amino acid alignments were largely uninformative. However, prior to creating the allelic swap of skaA as described above, we tried and failed numerous times to create a clean deletion of skaA in multiple strain backgrounds that were concurrently selected for megaplasmid maintenance. It is therefore possible that the SkaA protein is essential for proper replication and partitioning of megplasmid pMPPla107.

Combining comparative genomics, microbial genetics and evolutionary passage enabled us to identify two distinct genetic changes that affect sensitization of P. stutzeri to an inhibitory agent upon acquisition of genes [13]. We identify a region on pMPPla107 and a SNP in the gene we now call skaA that are responsible for resistance to the Pseudomonas inhibitory agent. Our data presented here are the framework on which to begin future work identifying the mechanism behind skaA and designing directed deletions within the 4B deletion that will be critical to identifying the other components regarding the inhibitory agent sensitivity phenotype associated with the acquisition of pMPPla107.

Data accessibility

We have included all raw data for the manuscript, either as accessioned sequencing files in the SRA or as electronic supplementary material files in Figshare. https://doi.org/10.6084/m9.figshare.15072537.v3.

Authors' contributions

B.A.S. conceived of the experiments, carried out molecular and bioinformatic work and wrote the manuscript. K.D. carried out experimental evolution passages. M.C. assisted with the preparation of materials and strain characterization. D.A.B. conceived of the experiments, helped to write the manuscript and carried out some bioinformatic analyses.

Competing interests

We declare we have no competing interests.

Funding

This work was supported through a grant from the United States National Science Foundation (NSF) IOS-1856556 to D.A.B.