Mutations in the lux Operon of Natural Dark Mutants in the Genus Vibrio

Abstract

Bacterial bioluminescence can display a wide range of intensities among strains, from very bright to undetectable, and it has been shown previously that there are nonluminous vibrios that possess lux genes. In this paper, we report the isolation and characterization of completely dark natural mutants in the genus Vibrio. Screening of over 600 Vibrio isolates with a luxA gene probe revealed that approximately 5% carried the luxA gene. Bioluminescence assays of the luxA-positive isolates, followed by repetitive extragenic palindromic-PCR fingerprinting, showed three unique genotypes that are completely dark. The dark mutants show a variety of lesions, including an insertion sequence, point mutations, and deletions. Strain BCB451 has an IS10 insertion sequence in luxA, a mutated luxE stop codon, and a truncated luxH. Strain BCB494 has a 396-bp deletion in luxC, and strain BCB440 has a frameshift in luxC. This paper represents the first molecular characterization of natural dark mutants and the first demonstration of incomplete lux operons in natural isolates.

Bioluminescence is a biochemically well-understood, easily identifiable physiological marker (13, 24) which, in most known species, is under quorum-sensing control (20, 23). In order to be capable of bioluminescence, bacteria require a minimum of five genes arranged in an operon: luxCDABE. Most bioluminescent species have additional nonessential genes in the lux operon as well, including (among others) luxG (found in all species except Photorhabdus luminescens) and luxH (found in Vibrio harveyi) (6, 11, 35).

For more than a century, bacterial bioluminescence has called attention to itself by its obvious visual phenotype. It is only since the advent of molecular probing techniques that we have become aware of the potentially widespread occurrence of non-visibly luminous strains (19, 27, 29). Palmer and Colwell (27) found a wide range of luminescence intensity levels among natural Vibrio cholerae isolates and discovered that a significant fraction of “visually non-luminescent” isolates contained a luxA gene. Lee and Ruby (19) took advantage of molecular probing to demonstrate that near-shore Hawaiian seawater contains a large number of non-visibly luminous Vibrio fischeri strains, consistent with the phenotype of the light organ symbionts of the bobtail squid Euprymna scolopes, which do not visibly glow under typical laboratory culture conditions but glow brightly in the squid. Using a V. harveyi luxA gene probe, Ramaiah et al. (29) found that an average of 1.7% of their Chesapeake Bay isolates had luxA but were not visibly bioluminescent. They suggested that either lux genes were not being expressed or that alterations or deletions had occurred in lux genes. The molecular basis of these dark phenotypes in natural populations has been, until now, unexplored.

Our work shows that natural populations of vibrios in Boca Ciega Bay, Florida, also include dark strains that contain lux genes. Additionally, we have discovered the molecular basis for the dark phenotypes of our isolates. In this paper, we report the detection of mutations in the lux operon of dark strains of Vibrio species. Screening of over 600 environmental Vibrio isolates from Boca Ciega Bay, Florida, revealed three dark genotypes. Sequence analysis of lux operons from the dark mutants reveals that one strain has an IS10 insertion sequence in luxA, a mutated luxE stop codon, and a truncated luxH gene. A second mutant has a deletion in luxC, while a third mutant has a frameshift in luxC. To our knowledge, this is the first molecular characterization of natural dark mutants.

MATERIALS AND METHODS

Sampling site and collection of planktonic isolates.

Water samples used in this study were collected from Boca Ciega Bay, Florida. Sampling was done once a day from 27 September 2005 through 30 September 2005. Vibrios were isolated by spreading seawater onto TCBS (thiosulfate-citrate-bile salts-sucrose) agar plates (Difco) and then incubating the plates overnight at room temperature. Isolates were replated onto nutrient seawater complete (SWC) agar plates (3 ml of glycerol, 1g of yeast extract, 3g of peptone, and 15 g of agar per liter of 75% seawater) for transfer back to the laboratory. Bacterial strains used in this study are listed in Table 1.

TABLE 1.

Bacterial strains used in this study

Strain	Source	Relevant feature
BCB440	This study	Dark
BCB451	This study	Dark
BCB494	This study	Dark
BCB440/pEOG	This study	Bright
BCB451/pEOG	This study	Bright
BCB494/pEOG	This study	Bright
BCB440/pVSV105	This study	Dark
BCB451/pVSV105	This study	Dark
BCB494/pVSV105	This study	Dark
V. harveyi B392	ATCC 33843	Bright

PCR.

PCR was done using standard reaction mixtures from Promega (Madison, WI) in a Perkin Elmer 480 DNA Thermal Cycler using the following conditions: 5 min of denaturation at 94°C, followed by 30 cycles of 1 min at 94°C, 1 min at 5°C below the thermal denaturation midpoint temperature, and 1 to 2 min of extension at 72°C, depending on the target length. The PCR was terminated after a final extension of 7 min at 72°C. PCR primers used in this study are listed in Table 2.

TABLE 2.

Oligonucleotide primer and probe sequences used in this study

Primer or probea	Sequence (5′→3′)	Tm (°C)b
127 Fc	GARCAYCAYTTYACWGARTTTGG	53
1007 Rc	ATTTCWTSYTCAGWDCCRTTHGCTTCRAAWCC	60
BOX	CTACGGCAAGGCGACGCTGACG	53
VhluxC1 F1d	AACCCAGATGGTGAATCATG	53
VhluxC1 R1d	GGTTTGCTCTATTTCGACAG	51
VhluxC F2	GCAGCGTTGGAAGAGCGAAGA	60
VhluxC F3	ATGCAATGCGCCCTCTTCAACG	61
VhluxD F1	GACTATTGCACACGTGTTGCG	57
VhluxD F2	CGTCGACTTTATCGAGCCTGA	57
VhluxD R	CTAAGCCATTTCTGGCGTAC	54
VhluxA F3	CACTTA TCAGCCACCTGAGCT	57
VhluxA F4	GTAGGGACGCCTGAAGAGTGTATCGC	62
VhluxB F1	TCAAAGCGTTCTTCTGATCA	52
VhluxB F2	GCCGCGGACCTATTGATGTC	58
VhluxB R	GAGTGGTATTTGACGATGTTGG	54
VhluxE F1	ATGGACGTACTTTCAGCGGT	56
VhluxE F2	CCACCGTTGAGATCGTTAGAAGA	56
VhluxE R	TCAGTTACCTCCGTCATTCTTAGC	56
VhluxG F2	GCCGTTAACTAGCTTCATATTCCG	56
VhluxG F3	TCTACTTGTGTGGTCCTTAC	51
VhluxG R	CAGATAAGCGAACGCATCCG	56
VhluxH F1	ATGAGCTCAACGTCACTACT	53
VhluxH R1	CTAAGACCAACTCACTTCACGCA	57
VhUS F1	ACCTGTAACCAAATTGACGC	53
VhUS F2	GGCGTGGCTTGAGCATACG	59
VhUS F3	GACCGCCACACAACTATCAG	56
VhUS R1	TAGAGTGTTATGTGAGGGCT	53
VhUS R2	GATTACGATGATTATCGCCA	50
VhUS R3	CTGATAGTTGTGTGGCGGTC	56

^aF, forward primer; R, reverse primer; P, probe.

^bT_m, thermal denaturation midpoint temperature.

^cPrimers modified from Wimpee et al. (36).

^dPrimers designed based on sequence from V. harveyi B392.

Labeling of hybridization probes and colony hybridizations.

PCR products were purified through Qiagen PCR cleanup columns (Valencia, CA) and labeled with [α-³²P]dCTP (Perkin-Elmer Life Sciences, Wellesley, MA) using the Prime-A-Gene system purchased from Promega (Madison, WI). For initial luxA hybridizations, environmental Vibrio strains were inoculated individually onto gridded circular nylon transfer membranes (Osmonics) on SWC agar plates. The SWC plates were incubated for 16 h at room temperature, after which membranes were removed from SWC plates using sterile forceps, and colonies were lysed according to Sambrook et al. (32).

For screening of genomic libraries, colonies were transferred to membranes by colony lift. Membranes were prehybridized and hybridized using standard protocols (32).

DNA extraction and fingerprint analysis.

Genomic DNA was isolated from SWC broth cultures using the cetyltrimethylammonium bromide procedure (1). Bacteria were genotyped by repetitive extragenic palindromic PCR genomic fingerprinting using a BOX primer, as described by Rademaker and deBruijn (28). Gel images were digitally captured using a Gel Logic 100 Kodak Imaging System and examined visually for differences in genomic profiles of different strains.

Screening environmental clones for luminescence.

Environmental isolates showing positive luxA hybridization were stabbed onto SWC plates and grown overnight at room temperature. Colonies were then taken into a dark room and visually examined for bioluminescence. Environmental isolates that appeared to be visually dark were then stabbed onto a single SWC plate, placed on X-ray film in a light-protected box, and exposed overnight.

Genomic cloning of strains BCB440, BCB451, and BCB494.

Total genomic DNA was subjected to either partial Sau3A digestion (BCB451) or total Xba1 digestion (BCB440 and BCB494), followed by size selection of 10-kb fragments on 0.75% agarose gels. DNA was eluted from gels using Qiaex II (Qiagen) and ligated into pGEM3Z cloning vector (Promega, Madison, WI). Ligated plasmids were then moved into Escherichia coli strain XL10-Gold (Stratagene) by transformation. The transformed cells were plated on LB plates containing 100 μg/ml ampicillin, 50 μl of X-Gal (5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside; 50 mg/ml), and 50 μl of 100 mM IPTG for blue/white screening.

Plasmid isolation and DNA sequencing.

Plasmids were isolated using a QIAprep Spin Miniprep Kit (Qiagen, Inc.) and insertion ends were sequenced using primers specific to the T7 and SP6 promoter regions. Automated sequencing was done by the Cancer Research Center DNA Sequencing Facility at the University of Chicago. Internal lux operon sequences were generated using primers based on V. harveyi strain B392. Flanking regions were sequenced using primers based on the previous sequence runs (i.e., primer walking). To identify protein coding genes, DNA sequences were subjected to BLASTX analysis (12; http://www.ncbi.nlm.nih.gov) using the default parameters.

Genomic cloning of the V. harveyi B392 lux operon for complementation.

The lux operon from V. harveyi strain B392 (ATCC 33843) including regulatory elements was cloned into a KpnI-digested Vibrio shuttle vector, pVSV105 (8) (gift of Erik Stabb), which contains a chloramphenicol resistance cassette. The shuttle vector with insert was designated pEOG. pEOG was transformed into E. coli strain DH5αλpir (Erik Stabb, Active Motif, Carlsbad, CA) and plated on LB plates containing 25 mg/ml chloramphenicol, 50 μl of X-Gal (50 mg/ml), and 50 μl of isopropyl-β-d-thiogalactopyranoside (100 mM) allowing for blue/white screening.

Conjugation.

Triparental matings were carried out using E. coli DH5αλpir containing the pEOG as the donor, E. coli CC118λpir containing the conjugation helper plasmid pEVS104, and, in separate conjugations, BCB440, BCB451, or BCB494 as the recipients. Three empty vector controls, BCB440/pVSV105, BCB451/pVSV105, and BCB494/pVSV105, were also generated using triparental mating.

Light and turbidity measurements.

Strains BCB440/pEOG, BCB451/pEOG, BCB494/pEOG, BCB440/pVSV105, BCB451/pVSV105, BCB494/pVSV105, BCB440, BCB451, and BCB494 were inoculated and grown for 16 h at room temperature. After 16 h, samples were diluted 1:100, and luminescence and optical density readings were taken every hour over an 8-h time period. Luminescence was measured using a Lumac/3m Biocounter M2010 luminometer. Optical density at 600 nm was measured in all strains prior to light measurements on a Versamax Spectrophotometer plate reader.

Nucleotide sequence accession numbers.

The following sequences obtained in this study for the lux operon and flankers of the indicated strains were deposited in the GenBank database (accession numbers are in parentheses): BCB440 (EU192082), BCB451 (EU192083), and BCB494 (EU192084).

RESULTS

Hybridization of luxA DNA probe to environmental Vibrio isolates.

To detect dark phenotypes that possess lux genes, 610 environmental Vibrio isolates were screened with a mixed luxA probe from V. harveyi, V. fischeri, and Photobacterium leiognathi PCR products (25, 36). For hybridization controls, 21 previously identified visually luminous V. harveyi strains were also inoculated onto the nylon membrane grids. Of the 610 isolates, 30 hybridized with the mixed luxA probe (Fig. 1). Bioluminescence assays of the luxA-positive isolates, followed by repetitive extragenic palindromic PCR fingerprinting, showed three unique genotypes that are completely dark. Of the three genotypes, one was encountered twice (represented by strains BCB494 and BCB584) in separate samples from separate locations in Boca Ciega Bay. Because the two appear to have identical genotypes, BCB494 was chosen for further analysis. Based on luxCDABEGH, 16S, and recA sequences, we have tentatively identified BCB440 and BCB451 as V. harveyi. BCB494 is closely related to V. harveyi based on the 16S and luxCDABEGH sequences; however, its recA sequence identifies BCB494 as Vibrio cincinnatiensis.

FIG. 1.

Autoradiogram indicating the Vibrio isolates positive for luxA. Thirty of 610 isolates hybridized with a V. harveyi luxA gene probe. Control strains are shown at the bottom right.

PCR amplification and nucleotide sequencing of luxA and luxC DNA in dark strains.

In an attempt to detect anomalies that might explain the observed phenotypes, we amplified the lux genes from the dark mutants. Anomalous sizes were found for the luxA PCR product of BCB451 and for the luxC product of BCB494 (Fig. 2). PCR products from other lux genes were of the expected normal sizes (data not shown). Anomalous PCR products were sequenced to reveal the basis for their unexpected sizes.

FIG. 2.

Amplified PCR products obtained using luxA primers 127 F and 1007 R (left) and luxC primers VhluxC F1 and VhluxC R1 (right) from dark mutants and V. harveyi B392 (ATCC).

In strain BCB451, the anomalous luxA PCR product is explained by an IS10 insertion sequence that has invaded the luxA gene (Fig. 3A). The 1,351-bp IS10 insertion sequence begins 446 bp downstream from the luxA start codon and is flanked by 9-bp direct repeats, characteristic of IS10 (Fig. 4A and B). It is oriented in the direction opposite to that of the lux operon and displays IS10-type inverted repeats immediately flanking the direct repeats. For comparison, we aligned the BCB451 IS10 with that of Shigella flexneri. The IS10 transposon-translated sequence is 79% similar to that to that of S. flexneri. Alignment of the inverted repeats of S. flexneri and BCB451 showed a sequence identity of 20 out of 29 nucleotides (Fig. 4C). In addition, one of the inverted repeats has a 19-bp insertion relative to the sequence the of S. flexneri inverted repeat.

FIG. 3.

Gene map of BCB451 (A), BCB494 (B), and BCB440 (C) lux operons. The genes, insertion elements, and deletions are drawn to scale and positioned in the direction in which they are transcribed. Positions of point mutations in strains BCB451 and BCB440 are indicated. UTR, untranslated region.

FIG. 4.

Direct and inverted repeats of BCB451. (A) Schematic of the positions of direct and inverted repeats flanking IS10. (B) Sequence of direct repeats. (C) Alignment of BCB451 and S. flexneri inverted repeats.

In dark strain BCB494, sequence analysis of the luxC gene revealed that the first 396 bp of the normally 1,308-bp structural gene are deleted (Fig. 3B). The larger than expected luxC PCR product (Fig. 2) was a puzzle until sequencing revealed that the forward primer had fortuitously annealed to a region upstream of luxC with partial sequence identity. Immediately adjacent to the luxC deletion in BCB494 is a region containing truncated gene fragments with homology to two insertion sequences, IS222 and IS3, as well as a 353-bp AT-rich region. Both the IS3 and IS222 insertion sequences are oriented in the same direction as the lux operon and appear truncated at both the 5′ and 3′ ends, relative to their closest known homologs. The truncated IS222 insertion sequence is positioned 711 bp upstream of luxC and is approximately 144 bp in length. A functional IS222 is on average 840 bp in length, indicating that the IS222 fragment found in BCB494 appears to be missing 696 bp (Fig. 5A). No identifiable direct or inverted repeats were detected. The IS3 homolog is located 70 bp downstream of the IS222 fragment and is 147 bp in length. IS3 has been documented in two other vibrios, Vibrio parahaemolyticus and Vibrio vulnificus. In these two species, IS3 is 282 bp in length. Alignment of the IS3 sequences of V. parahaemolyticus, V. vulnificus, and BCB494 reveals that the BCB494 IS3 is missing the first 9 codons on the 5′ end and is truncated by 41 codons at the 3′ end (Fig. 5B). The BCB494 IS3 is flanked downstream by a 353-bp AT-rich region that is apparently untranslated, indicated by numerous stop codons in all three reading frames.

FIG. 5.

(A) Translated sequence of BCB494 truncated IS3 insertion sequence aligned with insertion sequences of V. parahaemolyticus (Vp) and V. vulnificus (Vv). (B) IS22 insertion sequence aligned with that of Pseudomonas aeruginosa.

The dark mutant BCB440 did not show any PCR anomalies. However, sequence analysis of the BCB440 lux operon revealed the insertion of a single base pair in the luxC gene, resulting in a frameshift (Fig. 3C). The inserted base pair is 693 bp downstream of the luxC start codon.

Although the mutations described above are sufficient to account for the dark phenotypes observed in strains BCB440, BCB451, and BCB494, the entire lux operon of each mutant was cloned and sequenced to determine whether additional debilitating mutations can be detected. No additional lesions were found in BCB494 and BCB440, but there were two additional mutations in BCB451. The luxE stop codon of BCB451 has mutated to TCA, and the luxH gene is truncated by 203 bp (Fig. 3A).

Complementation of dark mutants.

The identification and nature of the mutations in the three dark mutants raised the question as to whether genetic complementation would rescue the light phenotype. To test this, the entire lux operon of V. harveyi strain B392 was cloned and introduced into each of the three dark mutants by conjugation. As shown in Fig. 6, complementation with a functional lux operon restored light production in all three mutants. In addition, the rescued light-producing phenotype displayed the cell density dependence typical of quorum sensing.

FIG. 6.

(A-C) Luminescence measurements of dark mutants complemented with the lux operon and upstream regulatory region. (A) BCB440/pEOG (circles), BCB440/pVSV105 (triangles), and BCB440 (squares). (B)BCB451/pEOG (circles), BCB451/pVSV105 (triangles), and BCB451 (squares). (C) BCB494/pEOG (circles), BCB494/pVSV105 (triangles), and BCB494 (squares). (D-F) Quorum-sensing-regulated luminescence. Specific luminescence was calculated by dividing luminescence by the absorbance for each time point of strains BCB440/pEOG (D), BCB451/pEOG (E), and BCB494/pEOG (F).

DISCUSSION

It is well established that when any of the five essential lux operon genes, luxCDABE, are disrupted, light production is abolished (3, 9, 10). In this study, therefore, we focused our attention on the lux operon itself. We have found that each of our dark mutants shows a different type of impairment, deactivating one of the five canonical lux genes.

Consistent with the abnormally large luxA PCR product of strain BCB451, we find that its luxA gene has been invaded by an IS10 insertion sequence. Although the BCB451 insertion has the signature inverted repeats of IS10, there is an additional 19-bp insertion within the inverted repeat on the 5′ end of the transposase gene. Whether this impairs the function of the IS10 is unknown. It is possible that the insertion within this inverted repeat happened subsequent to transposition, and the insertion element is now immovably trapped in the luxA gene. In BCB494, the first 396 bp of luxC are deleted. The sequences adjacent to the luxC deletion are homologous to the known bacterial insertion elements IS3 and IS222. However, these insertion sequences have themselves suffered catastrophic deletions and are unlikely to be functional. Unlike strains BCB451 and BCB494, which immediately showed PCR anomalies that led us to discover their respective lesions, PCR analysis of BCB440 showed no obvious abnormalities. When we sequenced the operon, the only lesion we found was a single base pair frameshift in luxC. Complementation of all three dark mutants restored cell density-dependent light production, indicating that the lesions in their lux operons solely account for the dark phenotype of these strains. Had the complementation not restored light production, other physiological or regulatory processes might have been implicated, in addition to the structural mutations.

There is a precedent for insertion sequence invasion of lux genes in a recent report of transposase disruption of the luxF gene in a Photobacterium mandapamensis strain (14). Although the P. mandapamensis insertion does not knock out luminescence (since luxF is not essential for bioluminescence), there are, nevertheless, indications that the insertion might affect the intensity of light emission in that strain. Since the P. mandapamensis strain carrying this insertion in luxF was isolated as a light organ symbiont, this knockout is a natural experiment in selection on the light-producing phenotype.

Because the operon functions as an integrated system, erosion of other lux genes in these dark mutants might be expected. Only BCB451 shows noticeable deterioration of the operon, which occurs in two additional places. The mutated luxE stop codon of BCB451 would result in a frameshifted fusion of luxE and luxG. An additional abnormality in BCB451 is a truncated luxH gene. However, luxH is a homolog of the ubiquitous ribB, which would provide the compensatory function in the riboflavin synthesis pathway which is required for luminescence. We have no way of knowing whether these additional mutations occurred before or after the IS10 insertion in luxA.

Dark variants have long been known in the bacterial bioluminescence field (15, 22, 33). Such variants occur most often in old cultures, especially in broth. However, such variants generally arise as a subset of luminous populations and are often unstable and prone to reversion, implying a mechanism akin to phase variation. In contrast, our mutants were dark as primary colonies and have remained so in all subsequent cultures derived from them. The luxA hybridization assays were performed after one transfer, and DNA was isolated for PCR and cloning from the same colonies as the hybridization assays. For these reasons, we are confident that the mutations did not arise in the laboratory. Further evidence to support a natural origin of our mutants includes the following: (i) we encountered the BCB494 genotype twice (BCB584 has the same genotype), in separate samples from separate locations in Boca Ciega Bay; (ii) BCB451 has three different types of mutations within the lux operon, a circumstance that would be highly unlikely to occur simultaneously in the primary colony; and (iii) in BCB440, the single point mutation leading to a frameshift would have to have occurred in the single cell that gave rise to the primary dark colony (i.e., in the seawater sample itself).

It is generally assumed that the preservation of gene sequences is the result of positive selection, but selection in the lux operon has long been the subject of debate (2, 4, 5, 23, 34). Several Vibrio and Photobacterium strains trade light for nutrients while participating in symbiotic relationships with various fish and/or squid species (7, 23, 30, 31). The selective advantage for light production is obvious in these cases. However, the vast majority of luminous bacterial isolates (including ours) are planktonic. At any time of the year, Boca Ciega Bay water yields dozens of luminous isolates per milliliter, most of which have different genotypes (16, 17). By sheer number and diversity, it is clear that planktonic bioluminescent strains are very common, yet the selective pressure that maintains this process is not obvious. It has been suggested (15; discussed in reference 34) that dark mutants would out-compete otherwise isogenic bright strains, simply because of the cost of light production. However, expression of the lux operon in the planktonic state should normally be low (and therefore inexpensive), because it is under quorum-sensing regulation (23). Quorum sensing requires high population density, which should be very rare in planktonic forms (however, see references 18, 21, and 26). This regulatory requirement argues that under typical environmental conditions, the lux operon would be all but superfluous in planktonic forms, so its loss should not be disadvantageous. However, although our data show that dark mutants are not uncommon, they are in the minority of the strains that possess lux genes. This argues against the idea that bioluminescence is an expensive remnant that is on its way to extinction in all but the symbiotic forms or an unnecessary vestige that can be lost with no disadvantage.

To our knowledge, the deletions we have found in BCB494 luxC and BCB451 luxH represent the first report of incomplete lux operons in natural isolates. We predict that there are many more. It is likely that we are underestimating the number of dark mutants in our sample group, since our initial screening used only luxA as a probe. It would not be surprising to find other mutants that lack luxA but possess other lux genes or that have lost the lux operon entirely. Such strains have thus far escaped detection, simply because we have not yet looked for them. We are almost certainly seeing a snapshot of a very dynamic process.