Recombination between the dfrA12-orfF-aadA2 Cassette Array and an aadA1 Gene Cassette Creates a Hybrid Cassette, aadA8b

Abstract

Homologous recombination between closely related gene cassettes, such as aadA1 and aadA2, which are 89% identical, can create hybrid cassettes and hybrids of existing cassette arrays. A new cassette array, dfrA12-orfF-aadA8b, which was created by such a recombination event occurring within the aadA2 cassette in the dfrA12-orfF-aadA2 array, has been identified.

Gene cassettes are a major source of the resistance genes found in clinical, commensal, and environmental isolates of bacteria that are resistant to antibiotics. Most commonly, they are found in association with class 1 or class 2 integrons (3, 13). One growing group of gene cassettes encodes aminoglycoside (3′) (9) adenylyltransferases that confer resistance to both streptomycin and spectinomycin (9). The genes and the cassettes, which are named after the genes, are designated aadA with an Arabic numeral to distinguish distinct genes, namely, those that differ by at least 2% in both the DNA and protein sequences. Two of these cassettes, aadA1 and aadA2, were present in two of the earliest-known plasmids that confer resistance to multiple antibiotics, namely, NR1, also called R100 (7), and pSa (1, 14), and remain very common in modern-day isolates. Though they have both been found in various contexts, they recur in a few specific cassette arrays, e.g., aadA1 or aadA2 alone, oxa1-aadA1, dfrA1-aadA1, and dfrA12-orfF-aadA2, which appear to have become globally disseminated. These two cassettes have now been sequenced many times, leading to the identification of several variant sequences for each of them, each containing one or a few single base changes (Fig. 1; also Fig. 3 in reference 9). The high level of similarity between the aadA1 and aadA2 cassette sequences (89.3% DNA identity; cassette length, 856 bp) means that they share many stretches of sequence identity that allow homologous recombination between them to occur. Several hybrids between aadA1 and aadA2 that presumably arose by homologous recombination have already been reported (Table 1). By combining existing knowledge of variant sequences and of cassette arrays, it is potentially possible to track the movement of specific sets of gene cassettes within bacterial populations and identify events that have been involved in their creation and dissemination.

FIG. 1.

Variants of the aadA2 gene cassette. The linear form of the 856-bp aadA2 cassette is shown, with the 59-base element (59-be) represented by a filled box and an arrow below indicating the extent of the aadA2 gene. The bases found in the prototype aadA2 sequence (GenBank accession no. X68227) at positions that can vary are shown in the box, with their positions in the cassette sequence given above. Position 1 is the first T residue of the 1R site (GTTAGAC) of the 59-be. For simplicity, only changes found in at least two of the available aadA2 sequences are included and variations in the 59-be are not shown. Variant sequences are grouped according to base changes, with any corresponding amino acid changes indicated. Variants shown below the box are found in the dfrA12-orfF-aadA2 cassette array (accession numbers in Table 2). Variants found in other arrays are listed above the box, with accession numbers shown, except for T17A C216T variants found in GenBank accession no. AY263740, *AF458082, *AF486817, and *AY681136 (an asterisk indicates that the sequence contains additional unique changes). Seventeen further sequences are identical to the prototype, and seven more have additional differences found only in these entries.

TABLE 1.

Hybrids of aadA1 and aadA2 gene cassettes

GenBank accession no.	Namea	Original name	Order	Crossover positionb	Change(s) from standard cassettec		Cassette array	Species	Reference or source
					aadA2d	aadA1d
AF047479	aadA3	aadA3	2/1	758-760		C759T	aadA3-aacA1/orfG-orfsH to M-catB3
AF326210e	aadA8	aadA8	2/1	550-600		C759T + 4f	Unknown	Klebsiella pneumoniae	12
AY139603	aadA8b	aadA8	2/1	602-647	A402Gg	C759T	aadA8b	Uncultured	15
AY852272	aadA8b		2/1	602-647	T17A C216T		dfrA12-orfF-aadA8b	Mixed culture	This work
AY232671	aadA1	aadA1	2/1	88-101	T17A	C759T	aadA2/1-blaP1	Pasteurella multocida	16
AY171244e	aadA21	aadA21	2/1	148-159		GC634-635CGg	aadA21	Salmonella enterica serovar Typhimurium	2
AY550883	aadA21	aadA22	2/1	148-159		G790Ag	aadA21	Salmonella enterica serovar Choleraesuis
AJ809407	aadA21	aadA23	2/1	148-159		G444Ag	aadA21	Salmonella enterica serovar Agona
AB189176	aadA21	aadA23b	2/1	148-159		A529Gg	aadA21	Escherichia coli
						G790Ag
AY665771h	aadA12	aadA12	2/1/2	148-159			Unknown	Escherichia coli
				550-600
AB107663	aadA1	aadA1	1/2	783-786		G732C C759T	aadA1/2	Vibrio cholerae

^aCassettes with high identity (>98%) to aadA1 or aadA21 are named aadA1 or aadA21, respectively.

^bFor boundary x-y, the last base identifiable as matching only the “front” cassette is x − 1, and the first base identifiable as matching only the “back” cassette is y + 1.

^cUnless otherwise indicated, these differences are found in a number of other variants of the relevant cassette.

^dThe standard cassettes used for comparison are found in GenBank accession no. X12870 for aadA1, variants of which commonly have C759T, and X68227 for aadA2.

^eThe first 9 nucleotides of the cassette sequence are missing from AF326210, and the last 28 bp of the cassette sequence are missing from AY171244.

^fThe four additional changes are C609A, A612G, G773A (also found in the aadA1 cassette variant in AY139597), and the unique change C599T.

^gThe changes indicated have not been seen in any other variant of the relevant cassette.

^hThe sequence of the aadA12 gene only is present in AY665771.

We used primers located in the 5′- and 3′-conserved segments (CS) of class 1 integrons (HS458, 5′-GTTTGATGTTATGGAGCAGCAACG-3′, and HS459, 5′-GCAAAAAGGCAGCAATTATGAGCC-3′ [5]) to screen DNA isolated from mixed bacterial samples recovered from the feces of human volunteers with no recent exposure to antibiotics. Mixed cultures were grown in L broth at 37°C under aerobic conditions, and the organisms recovered were predominantly Escherichia coli. DNA was recovered using alkaline lysis, and PCR conditions were as follows: denaturation at 94°C for 3.0 min; 35 cycles at 94°C for 30 s, 65°C for 1.0 min, and 72°C for 1.5 min or 4.0 min in the final cycle. A significant proportion of the samples screened yielded an amplicon of 2.2 kb corresponding to a cassette array of 1.7 to 1.8 kb when corrected for the amplified portions of the 5′-CS and 3′-CS (465 bp). Partial sequencing of seven of the PCR products revealed that the cassette array was either dfrA12-orfF-aadA2 (in one case) or a derivative in which the end of the aadA2 cassette is replaced by the corresponding part of the aadA1 cassette (in six cases) (Fig. 2). The 2.2-kb amplicon from one of the latter samples was cloned into pGEM-T Easy (Promega) by following the manufacturer's instructions, with selection on LB plates containing streptomycin (25 μg ml⁻¹) and trimethoprim (50 μg ml⁻¹), and the resultant plasmid (pMAQ697) also conferred resistance to spectinomycin (25 μg ml⁻¹). The insert was sequenced using procedures described previously (11). The crossover in the hybrid aadA2/aadA1 cassette was located between positions 602 and 647 in the cassette (numbered from the conserved TT at beginning of the cassette) by comparison with the reference aadA1 (GenBank accession no. X12870) and aadA2 (X68227) cassette sequences. A similar hybrid with the same crossover position has recently been reported as aadA8 (AY139603 [15]). However, the original aadA8 cassette (AF326210 [12]), which is also an aadA2/1 hybrid, has a different crossover position, between 550 and 600 (Table 1). As the three sequences exhibit high levels (>98%) of DNA and protein identity, we have named the aadA2/1 hybrid cassettes with crossover between positions 602 and 647 aadA8b (Table 1) to indicate the difference in the crossover position.

FIG. 2.

Schematic of the original and hybrid cassette arrays. The aadA8b cassette has arisen by homologous recombination between the aadA2 cassette (hatched) in a dfrA12-orfF-aadA2 cassette array (e.g., GenBank accession no. AF284063) and an aadA1 cassette (shaded) in another integron. Cassettes are represented as narrow open boxes with a larger box at one end representing the 59-base element and are shown to scale. × indicates the position of the crossover. The region (between positions 602 and 647) where recombination took place is indicated by a small filled box. The attI1 site is shown as a tall open box and defines the end of the 5′-CS (thick black line). The thin black line beyond the last cassette represents the 3′-CS sequence.

To examine if the recombination event could have occurred between the dfrA12-orfF-aadA2 cassette array in one integron and an aadA1 cassette located in a second integron, the sequences were examined in detail. The dfrA12-orfF-aadA2 cassette array was first isolated from an E. coli strain from Finland (4), but the sequence of the aadA2 cassette was not completed (GenBank accession no. Z21672). However, this array has since been found in many bacterial species (Table 2) isolated in many different countries from clinical and animal-associated sources and from wastewater. Examination of the sequences of this array recorded in the relevant GenBank entries revealed that five are identical to one another (reference sequence AF284063) and to the incomplete Z21672 sequence. The remainder differ by only a few single base changes (Table 2; Fig. 1), and some of these variations may represent either sequence errors or errors arising during PCR amplification. This suggests that this array arose on a single occasion and has become globally disseminated but subsequently acquired occasional base substitutions.

TABLE 2.

dfrA12-orfF-aadA2 cassette arrays

GenBank accession no.	Sequence difference(s) in cassettea			Organism	Reference or source
	dfrA12	orfF	aadA2
Z21672b			—c	Escherichia coli	4
AF284063d				Serratia marcescens
AB154407				Escherichia coli
AB191048				Staphylococcus aureus
AF550415				Citrobacter freundii
AY748453				Klebsiella pneumoniae
AY852272			—e	Mixed culture	This work
AB191047		T178C		Pseudomonas aeruginosa
AF175203	AAAA528-31AAAf	GG302-303GGGf		Citrobacter freundii
AY522923			G568A	Aeromonas hydrophila
AF188331	GGG101-103GGg		G568A	Shigella flexneri
AB196348	G300A G360A	G302Tf	A601G	Enterococcus faecalis
AY115474	C567Gf	C50G G75A T133C C140G T248C T267Gf	A601G	Uncultured	15
AY139605	C567Gf	C50G G75A T133C C140G T248C T267Gf	A601G	Uncultured	15
AY551331	T115C C179T		A732G	Salmonella enterica serovar Choleraesuis	6
AF335108			T216Ch CCG826-828Δf	Escherichia coli	8
AF180731		GC209-210CG	A34G G642T C826Δf G830Δf	Klebsiella pneumoniae

^aUnless indicated otherwise, these changes are seen only in GenBank entries included in this table.

^bStandard sequence for the dfrA12 and orfF cassettes.

^cSequence is only available to position 125 of the aadA2 cassette.

^dStandard sequence for the complete dfrA12-orfF-aadA2 array.

^eThe aadA8b hybrid cassette is identical to the aadA2 cassette in AF284063 to position 602 and to the standard aadA1 cassette in X12870 from positions 647 to 856.

^fIn the 59-base element.

^gThis change disrupts the dfrA12 reading frame, and there are likely errors elsewhere in this sequence (10).

^hFound in the prototype aadA2 sequence in X68227.

There are several variant sequences for the aadA2 cassette found among those deposited in GenBank (Fig. 1), but aadA2 in the dfrA12-orfF-aadA2 arrays (Fig. 1) usually differs at two positions (T17A and C216T) from the reference aadA2 cassette found in In6 from plasmid pSa (X68227). In the dfrA12-orfF-aadA2 and dfrA12-orfF-aadA8b arrays recovered here, the aadA2 gene cassette and the aadA2 portion of the aadA8b cassette are identical to the aadA2 variant found in the standard dfrA12-orfF-aadA2 array (AF284063), indicating that the recombination event is likely to have occurred within this array. Beyond the position of the switch to aadA1, the sequence of the aadA8b cassette is identical to the prototype for aadA1 (X12870) and to single aadA1 cassettes recovered from other samples in our commensal mixed-culture collection. In contrast, the variant of aadA8b found as a single cassette in AY139603 (15) differs from the version reported here at three positions in the aadA2 portion and one in the aadA1 portion (Table 1), indicating that the two variants are likely to have arisen via independent recombination events occurring within the same span (positions 602 to 647).

It is clear that recombination events involving the aadA1 and aadA2 cassettes have occurred on several occasions (Table 1) and thus contribute significantly to the creation of new cassette arrays. However, the fact that the dfrA12-orfF-aadA8b array was recovered from the fecal flora of a number of different individuals indicates that it has already become widely distributed, at least in Australia.

Nucleotide sequence accession number.

The sequence of the dfrA12-orfF-aadA8b PCR amplicon has been deposited in GenBank under accession no. AY852272.