Dissemination of IncI plasmid encoding bla _CTX-M-1 is not hampered by its fitness cost in the pig’s gut

ABSTRACT

Multiresistance plasmids belonging to the IncI incompatibility group have become one of the most pervasive plasmid types in extended-spectrum beta-lactamase-producing Escherichia coli of animal origin. The extent of the burden imposed on the bacterial cell by these plasmids seems to modulate the emergence of “epidemic” plasmids. However, in vivo data in the natural environment of the strains are scarce. Here, we investigated the cost of a bla _CTX-M-1-IncI1 epidemic plasmid in a commensal E. coli animal strain, UB12-RC, before and after oral inoculation of 15 6- to 8-week- old specific-pathogen-free pigs. Growth rate in rich medium was determined on (i) UB12-RC and derivatives, with or without plasmid, in vivo and/or in vitro evolved, and (ii) strains that acquired the plasmid in the gut during the experiment. Although bla _CTX-M-1-IncI1 plasmid imposed no measurable burden on the recipient strain after conjugation and during the longitudinal carriage in the pig’s gut, we observed a significant difference in the bacterial growth rate between IncI1 plasmid-carrying and plasmid-free isolates collected during in vivo carriage. Only a few mutations on the chromosome of the UB12-RC derivatives were detected by whole-genome sequencing. RNA-Seq analysis of a selected set of these strains showed that transcriptional responses to the bla _CTX-M-1-IncI1 acquisition were limited, affecting metabolism, stress response, and motility functions. Our data suggest that the effect of IncI plasmid on host cells is limited, fitness cost being insufficient to act as a barrier to IncI plasmid spread among natural population of E. coli in the gut niche.

INTRODUCTION

Since the 2000s, a global epidemic of extended-spectrum beta-lactamase (ESBL)-producing Enterobacterales has emerged, linked to the dissemination of successful bla _CTX-M-carrying clones (1). The CTX-M enzymes are mostly plasmid-encoded and E. coli is the major host (2).

Among major multiresistant plasmids in Enterobacterales, IncF, IncI, IncC, IncN, and IncL plasmid families are the most commonly reported. The I-complex plasmids contain incompatibility groups I, K, B, and Z. IncI, or MOB_P according to relaxase typing, is a group of low copy-number, narrow-host-range, conjugative plasmids, which vary in size from 50 to 250 kb (2, 3). Some indistinguishable IncI1 and IncIγ plasmids have been isolated from multiple, unrelated high-risk clones, identified in different periods and bacterial species, in distant geographical areas, defined as “epidemic” plasmids (4 – 7). IncI pST3 plasmid is the most commonly reported bla _CTX-M-1-IncI1 plasmid in Europe, described in pathogenic human E. coli isolates, and in E. coli and Salmonella spp. from food-producing animals (8). Moreover, Lucas et al. recently showed that resistance to extended-spectrum cephalosporins (ESC) in French pigs’ E. coli isolates is mainly carried by highly similar bla _CTX-M-1-IncI1/ST3 plasmids (9). IncI plasmids have been recognized in clinically relevant bacteria from human, animal, and environmental sources, and as major vehicles for the dissemination of ESBL, notably CTX-M-15 and CTX-M-1 enzymes (2, 8).

Previous studies using a series of experiments that included pigs inoculated with E. coli harboring bla _CTX-M-1-encoding IncI plasmid showed the maintenance of the plasmid in its E. coli host, in the pig’s gut. Moreover, the loss of the plasmid was a rare event in gut microbiota during these longitudinal studies of in vivo carriage in pigs (10). In these models, the transfer of the bla _CTX-M-1-encoding IncI plasmid between different E. coli strains occurred, underlining the dissemination capacity of the ESC resistance, independently of the exposure to antibiotic drugs (11, 12). Finally, understanding the factors of IncI1 plasmids success better is important, as they are major contributors to the dissemination of ESBL genes especially in the animal reservoir.

The arrival of a plasmid in a new bacterium produces a highly variable fitness cost depending on the plasmid–bacterium association (13 – 15). Plasmid persistence and dissemination in bacterial population, under conditions that do not select for plasmid-encoded genes, are influenced by the fitness effects associated with plasmid carriage (16, 17). Previous works have shown that compensatory adaptation frequently ameliorates the fitness of costly plasmids (18 – 20). This compensatory evolution may play a key role in the dynamics of the spread and evolution of antibiotic resistance plasmids (13, 21 – 23). However, most studies on the evolutionary dynamics of plasmid–bacterium associations have been performed in laboratory-based culture systems.

We took the unique opportunity of the previous studies performed in pigs for probiotic testing (12), to evaluate the fitness cost of IncI pST3 plasmid on bacterial isolates collected during an in vivo longitudinal study of carriage of a CTX-M-1-producing E. coli B1 commensal strain in pigs. We also investigated the impact of IncI pST3 plasmid on the bacterial host using whole-genome sequencing (WGS) and transcriptome sequencing and analysis to determine the effect of plasmid carriage on global gene expression.

RESULTS

Characterization of ancestral and in vivo evolved UB12 strains

According to the whole-genome sequencing results, E. coli UB12-RC recipient strain belongs to O100:H40 serotype, B1 phylogroup, and ST2520, according to the Warwick multilocus sequence typing (MLST) scheme. This strain harbors a mutation in rpoB after selection in Mueller−Hinton medium (MH) containing 250 mg/L rifampicin (10), and contains two plasmids: one P1-like phage-plasmid (IncY, 98,119 bp) (24), and one small mobilizable plasmid (Col156 plasmid, 5223 bp) (Table 1). In addition to the IncY and Col156 resident plasmids, the UB12-TC transconjugant harbors a bla _CTX-M-1-IncI plasmid (108699 bp), and one colicinogenic plasmid, belonging to the ColE1 group (ColRNAI plasmid, 6724 bp), which disseminated from the M63-DN donor strain. The bla _CTX-M-1-IncI plasmid is of plasmid multilocus sequence type 3 (pST3) of the IncI1 family and presents 99% identity and 72% coverage to R64 plasmid, a prototype of the IncI1 group (GenBank accession number AP005147) (8). Moreover, the core genome of IncI ST3 plasmid presents with other bla _CTX-M-1-IncI1/ST3 plasmid previously described 100% identity and 99% coverage (GenBank accession number KJ484635) (25). The bla _CTX-M-1 gene is associated with an ISEcp1 mobile element found at the 5’ end of the gene, and the ISEcp1-bla _CTX-M-1 transposition unit is inserted in the shufflon region of the plasmid. Other resistance genes carried in the bla _CTX-M-1-IncI/ST3 plasmid are dfrA17, sul2, and aadA5 (Table 1).

TABLE 1.

Characterization of donor E. coli strain, evolved and ancestral E. coli UB12 clones, and new in vivo transconjugants used in this study

Status	Isolate name	Experiment (month/year)	Box a	Animal a	Day of evolution	Phylogroup	MLST type b	Serotype	FimH allele	Plasmid replicon c	Antimicrobial resistance genes	Mobility d	Biofilm formation e (SD)
Ancestral strains	M63-DN		-	-	-	B1	ST2521	O147:H7	fimH31	IncI (KJ484635); ColRNAI (J01566)	bla CTX-M-1 ,sul 2, aadA5, dfrA17	+	0.079 (+/− 0.055)
	UB12-RC		-	-	-	B1	ST2520	O100:H40	fimH31	IncY (MH42254); Col156 (EU090225)	-	+	0.114 (+/− 0.066)
	UB12-TC		-	-	-	B1	ST2520	O100:H40	fimH31	IncI (KJ484635); ColRNAI (J01566); IncY (MH42254); Col156 (EU090225)	bla CTX-M-1 ,sul 2, aadA5, dfrA17	+/−	0.081 (+/− 0.048)
Evolved strains with IncI plasmid	UB12-TCive.1	January/2013	F3v	4368	D28	B1	ST2520	O100:H40	fimH31	IncI (KJ484635); ColRNAI (J01566); IncY (MH42254); Col156 (EU090225)	bla CTX-M-1 ,sul 2, aadA5, dfrA17	+/−	0.094 (+/− 0.047)
	UB12-TCive.2	June/2013	B2d	238	D8							+/−	0.092 (+/− 0.048)
	UB12-TCive.3	June/2013	B4v	4631	D31							+/−	0.091 (+/− 0.044)
	UB12-TCive.4	October/2012	B1	4219	D7							+/−	0.090 (+/− 0.048)
	UB12-TCive.5	October/2012	B1	4219	D20							+/−	0.092 (+/− 0.042)
	UB12-TCive.6	December/2015	C3d	462	D18							+/−	0.086 (+/− 0.043)
	UB12-TCive.7	December/2015	C4v	5652	D45							+/−	0.091 (+/− 0.053)
Evolved strains with IncI plasmid loss	UB12-TCΔpIncI.1	June/2013	B2v	4613	D20	B1	ST2520	O100:H40	fimH31	ColRNAI (J01566); IncY (MH42254); Col156 (EU090225)	-	+	0.100 (+/− 0.067)
	UB12-TCΔpIncI.2	April/2012	C3v	3716	D7							+	0.104 (+/− 0.054)
	UB12-TCΔpIncI.3	April/2012	C3v	3716	D18							+	0.095 (+/− 0.041)
New in vivo transconjugants with IncI plasmid	NI.1	January/2013	B2v	4331	D22	D	ST13022	-:H25	fimH123	IncI (KJ484635); IncY (MH42254); IncFII (pRSB107)	bla CTX-M-1, sul 2, aadA5, dfrA17	+	ND
	NI.2	January/2013	F3v	4324	D22	A	ST48	O184:H11	fimH41	IncI (KJ484635); IncFIC(FII), IncFIB (AP001918); IncFII (pSFO) (AF401292); ColE10 (X01654); ColRNAI (J01566)	bla CTX-M-1, sul 2, aadA5, dfrA17, sul1, tetA, tetB	+/−	ND
	NI.3	January/2013	F3v	4368	D10	B1	ST13023	-:H7	fimH68	IncI (KJ484635)	bla CTX-M-1, sul 2, aadA5, dfrA17	+	ND
	NI.4	June/2013	B2v	4591	D16	B1	ST2540	OgN9:H2	fimH31	IncI (KJ484635); ColRNAI (J01566)	bla CTX-M-1, sul 2, aadA5, dfrA17	+	ND
	NI.5	June/2013	B2v	4575	D13	G	ST657	O149:H25	fimH97	IncI (KJ484635); IncFII(pRSB107)(AJ851089); ColRNAI (J01566); Col(MG828)	bla CTX-M-1, sul 2, aadA5, dfrA17	+	ND
	NI.6	June/2013	B2d	238	D8	C	ST410	Onovel14:H9	fimH24	IncI (KJ484635); IncFIC(FII), IncFIB (AP001918), IncFII (pSFO) (AF401292); ColRNAI (J01566)	bla CTX-M-1, sul 2, aadA5, dfrA17, tetA	−	ND
	NI.7	October/2012	B1	4219	D20	G	ST117	O24:H4	fimH97	IncI (KJ484635); IncFIC(FII), IncFIB, IncF1A (AP001918), IncFII (pSFO) (AF401292); ColRNAI (J01566); Col(MG828)	bla CTX-M-1, sul 2, aadA5, dfrA17	+	ND
	NI.8	December/2015	C4d	5625	D4	B1	ST767	O8:H9	fimH32	IncI (KJ484635); IncFIC(FII), IncFIB (AP001918); IncFII (pSFO) (AF401292); Col(MG828)	bla CTX-M-1, sul 2, aadA5, dfrA17, tetA	+	ND
	NI.9	December/2015	C4d	5625	D11	B2	ST95	O18:H7	fimH15	IncI (KJ484635); IncFIC(FII), IncFIB (AP001918), IncFII (pSFO) (AF401292); ColRNAI (J01566)	bla CTX-M-1, sul 2, aadA5, dfrA17	+/−	ND
	NI.10	December/2015	C4v	5626	D11	G	ST117	-:H19	fimH97	IncI (KJ484635); IncFIB (AP001918), IncFII (pRSB107); Col(MG828); Col8282 (DQ995353); p0111; ColRNAI (J01566)	bla CTX-M-1, sul 2, aadA5, dfrA17, tetB	−	ND
	NI.11	December/2015	C4d	5651	D11	C	ST88	O8:H9	fimH39	IncI (KJ484635)	bla CTX-M-1, sul 2, aadA5, dfrA17	−/+	ND

^a Detailed in references (9, 11).

^b According to the Warwick multilocus sequence typing (MLST) scheme.

^c PlasmidFinder, IncI plasmid is in bold.

^d +, total spreading within medium; +/−, partial spreading; −, not spreading.

^e Mean of absorbance measure ancestral tranconjugants at 570 nm. SD, standard deviation; ND, not determined.

Three cefotaxime-susceptible and rifampicin-resistant isolates, collected at days 7, 18, and 20, during longitudinal carriage of the UB12-TC strain in the pigs, have spontaneously lost the IncI plasmid in vivo and harbor Col156, IncY, and ColRNAI plasmids (referenced as UB12-TCΔpIncI) (Table 1).

To identify mutations and rearrangements that occur during the carriage of UB12-TC strains in the pig’s gut, we used the open-source Breseq pipeline (26) and the ancestral UB12-TC strain as reference genome. In the first step, we excluded all single nucleotide polymorphisms (SNPs), insertions or deletions detected comparing mapping reads of the UB12-TC ancestral strain to the assembly of itself. In the second step, we determined genetic differences occurring between the UB12-RC recipient strain, seven UB12-TC isolates, collected in stools between day 7 and day 45 after oral inoculation of pigs, i.e., in vivo evolved (UB12-TC^ive), and the cefotaxime-susceptible and rifampicin-resistant isolates (UB12-TCΔpIncI) (Table 2). All transconjugants showed one intergenic point mutation, upstream of the nadA gene, except two (Table 2). In chromosomal DNA, we detected seven other events between the UB12-RC recipient strain, the UB12-TC strain and six UB12-TC^ive transconjugants: one intergenic point mutation between the ygeH and ygeG genes, one synonymous point mutation in the kdsB gene, and five nonsynonymous point mutations in the ecpC, ydeE, ttdB, thiH genes, and in the hlyE pseudogene (Table 2). In total, 1–3 point mutations per strain were detected. In plasmid DNA, a common rearrangement was detected in the shufflon region of the IncI plasmid in three transconjugants. None of other plasmids showed any modification.

TABLE 2.

Characterization of mutations and rearrangements in the chromosome and plasmids of ancestral and evolved E. coli UB12 clones

Strains	Mutations a		Rearrangements b
	Plasmids	Chromosome	Plasmids	Chromosome
UB12-RC (recipient)	ΔpIncI, ΔColRNAI	-	-	-
UB12-TC (ancestral transconjugant)	-	intergenic region (−286/–); quinolinate synthase gene (nadA)/ (T→A)	-	-
UB12-TCive.1	-	intergenic region (−286/–); quinolinate synthase gene (nadA)/ (T→A)	pIncI: –/Mobile element protein, ISEcp1/Tryptophan synthase beta chain like/IncI1 plasmid conjugative transfer pilus-tip adhesin protein PilV
		putative fimbrial usher protein EcpC (ecpC) / V340L (GTG→TTG)
UB12-TCive.2	-	intergenic region (−286/–); quinolinate synthase gene (nadA)/ (T→A)	-	-
UB12-TCive.3	-	intergenic region (−286/–); quinolinate synthase gene (nadA)/ (T→A)	-	-
UB12-TCive.4	-	-	pIncI: Tryptophan synthase beta chain like/IncI1 plasmid conjugative transfer pilus-tip adhesin protein PilV	-
UB12-TCive.5	-	intergenic region (−286/–); quinolinate synthase gene (nadA)/ (T→A)	pIncI: –/Mobile element protein, ISEcp1/Tryptophan synthase beta chain like/IncI1 plasmid conjugative transfer pilus-tip adhesin protein PilV	-
		intergenic region (−207/+128); putative transcriptional regulator (ygeH) / TPR repeat-containing putative chaperone (ygeG) / (T→C)
UB12-TCive.6	-	intergenic region (−286/–); quinolinate synthase gene (nadA)/ (T→A)	-	-
		dipeptide exporter (ydeE) / R23H (CGC→CAC)
		3-deoxy-manno-octulosonate cytidylyltransferase (kdsB) / R38R (CGC→CGA)
UB12-TCive.7	-	intergenic region (−286/–); quinolinate synthase gene (nadA)/ (T→A)	-	-
		L(+)-tartrate dehydratase subunit beta (ttdB) / E110K (GAA→AAA)
UB12-TCΔpIncI.1	ΔpIncI	intergenic region (−286/–); quinolinate synthase gene (nadA)/ (T→A)	-	-
UB12-TCΔpIncI.2	ΔpIncI	2-iminoacetate synthase (thiH) / A147S (GCC→TCC)	-	-
UB12-TCΔpIncI.3	ΔpIncI	intergenic region (−286/–); quinolinate synthase gene (nadA)/ (T→A)	-	-
		fragment of hemolysin E (pseudogene, hlyE) / (C→A)

^a Mutations and rearrangements are detected in the chromosome or plasmids of each strain compared to the UB12-RC genome.

^b Nucleotide changes are given in parentheses and underlined; synonymous mutations are in bold; the position of intergenic mutations is shown by the relative distance with start or stop codon of genes; Δ = deletion.

No fitness cost decrease of UB12 strains during digestive carriage in pigs

To evaluate the IncI plasmid cost on the bacterial fitness, we compared the maximum growth rates (MGRs) of the E. coli M63-DN donor strain, the UB12-RC recipient strain and the ancestral UB12-TC transconjugant. The MGR of M63-DN donor strain (3.55 div h-1) was significantly higher than the MGR of UB12-RC strain (3.38 div h-1) (P = 0.02). The difference between the MGR of UB12-RC strain and the MGR of UB12-TC transconjugant (3.28 div h-1) was nonsignificant (P = 0.09) (Fig. 1).

Fig 1.

Boxplot representation of the maximum growth rate (MGR) of each pool lineage: MGR of the donor strain (M63-DN), the recipient strain (UB12-RC), the ancestral transconjugant (UB12-TC, IncI, pST3), seven in vivo evolved transconjugants (UB12-TC^ive), and three evolved transconjugants which had lost in vivo the IncI plasmid (UB12-TCΔpIncI). UB12-TC^ive and UB12-TCΔpIncI were isolated in stools between day 7 and day 45 during longitudinal study of in vivo carriage in pigs. For each box, the central mark indicates the median, and the bottom and top edges of the box indicate the 25th and 75th percentiles, respectively. Each point corresponds to one measurement per clone, which were repeated six times per strains. Asterisks indicate significant differences (* =P < 0.05, ** =P < 0.01 and *** =P < 0.001). NS = not significant.

To evaluate the adaptation of IncI plasmid in the E. coli UB12 strain, we compared the MGRs of the UB12-TC strain, the M63-DN donor, and 10 clones (UB12-TC^ive and UB12-TCΔpIncI) isolated from the stools of the pigs, between day 7 and day 45, during an in vivo longitudinal study of carriage of the UB12-TC transconjugant (11). Among these clones, three UB12-TC transconjugants had spontaneously lost the IncI plasmid (UB12-TCΔpIncI). As no significant fitness cost variation was detected between the MGRs of the evolved UB12-TC^ive transconjugants on the one hand, and between the MGRs of the UB12-TCΔpIncI transconjugants on the other hand, further statistical analyses were performed after pooling the MGRs of the clones from each group. The difference between the MGRs of UB12-TC transconjugants (3.28 div h-1) and the MGRs of UB12-TC^ive evolved transconjugants (3.23 div h-1) was nonsignificant (P = 0.08), as the difference between the MGRs of UB12-RC clones and the MGRs of the UB12-TC^ive transconjugants (P = 0.07). Interestingly, the differences between the MGRs of UB12-TCΔpIncI (3.40 div h-1) and the MGRs of UB12-TC or UB12-TC^ive clones were significant (P = 0.004 and P = 1.61 × 10⁻⁷, respectively) (Fig. 1).

No fitness cost decrease during in vitro experiment evolution of UB12-TC

To assess if the absence of decrease in in vivo fitness cost could be reproduced in vitro in a constant well-defined environment, we conducted an in vitro evolution experiment in LB during 30 days (300 generations in total) and evaluated the impact of the IncI plasmid on the bacterial host after evolution. We measured the MGRs of five independent ancestral transconjugants (UB12-TC), and of five independent clones for each lineage after in vitro evolution (denoted UB12-TC^ite), compared to the MGRs of five plasmid-free ancestral and evolved control strains (UB12-RC and UB12-RC^ite). First, no significant fitness cost variation was detected between the MGRs of the evolved clones belonging to the same lineage. Further statistical analyses were performed after pooling the MGRs of the clones from each lineage. The differences between the MGRs of UB12-TC (3.49 div h-1) and UB12-TC^ite (3.35 div h-1) clones, and the MGRs of plasmid-free control strains UB12-RC and UB12-RC^ite (3.49 div h-1 and 3.42 div h-1, respectively) were nonsignificant (Fig. 2). Although no significant difference was found between ancestral and evolved strains, the MGRs of UB12-TC decreased after the evolution experiment (3.49 div h-1 versus 3.35 div h-1 for UB12-TC and UB12-TC^ite, respectively). Of note, there was a trend (P = 0.054) for increased fitness cost in evolved strains (Fig. 2).

Fig 2.

 Boxplot representation of the maximum growth rate (MGR) of each pool lineage: MGR of five ancestral UB12-RC and UB12-TC control strains and evolved UB12-RC and UB12-TC strains at day 30. For each box, the central mark indicates the median, and the bottom and top edges of the box indicate the 25th and 75th percentiles, respectively. Each point corresponds to one measurement per clone, which were repeated six times per transconjugant and control. NS = not significant.

Description of the newly transconjugants E. coli isolates (NI) with IncI plasmid

As the recipient host background could interfere in the spread of multidrug resistance plasmids, we analyzed the genetic diversity of 11 cefotaxime-resistant and rifampicin-susceptible newly formed transconjugants (natural isolates, NI) harboring IncI. The phylogroups of NI are diverse (one A, B2, and D, two C, three B1, and G) (Fig. 3A). Based on plasmid-families replicon, the plasmidome of NI appears varied with multiple combinations of plasmid types. Most of the eleven isolates (eight NI) carry Col-like plasmids, alone or co-carried with IncF plasmids. One isolate (NI-1) carries an IncY plasmid, sharing 98% identity and 71% coverage to the IncY plasmid of UB12-TC (Table 1). Eight NI (72%) harbor more than two plasmids, while two NI (NI-3 and NI-11) harbor the IncI plasmid alone (Table 1).

Fig 3.

(A) Phylogenetic tree of the 11 newly formed transconjugants detected among pig E. coli commensal isolates (NI clones, E. coli, IncI, pST3), the M63-DN donor strain, and the UB12-TC transconjugant strain (in bold) are presented. The tree is rooted on an Escherichia clade (E1492cladeI (27)). The phylogenetic tree was constructed with IQ-TREE v1.6.9 from the core genome genes using Roary v1.007002. From left to right are presented the Sequence Type (according to the Warwick University scheme), the O and H serotypes, the fimH allele, and the fitness cost of NI clones compared to the donor strain (M63-DN) (NA = not applicable, NS = not significant, * =P < 0.05, ** =P < 0.01, and *** =P < 0.001). Each branch of the same phylogenetic group is colored (yellow, D; red, B2; orange, G; blue, A; light green, C; dark green, B1). Information on O group, H type, and fimH allele are given in Table 1. (B) Boxplot representation of the maximum growth rate (MGR) of each strain : donor strain (M63-DN), UB12-TC ancestral transconjugant and NI clones, isolated in stools between day 4 and day 22 during longitudinal study of in vivo carriage in pigs. For each box, the central mark indicates the median, and the bottom and top edges of the box indicate the 25th and 75th percentiles, respectively. Each point corresponds to one measurement per clone, which were repeated six times per strains. Asterisks in red indicate significant differences (* =P < 0.05, ** =P < 0.01, and *** =P < 0.001) in comparison to M63-DN. Colors correspond to the strains of the same phylogroup (color code as in A).

The core genomic divergence, estimated by the number of core genome SNPs between strains, is variable and dependent on the sequence type and phylogroup of strains, from 18513 to 77175 (between E. coli B1 ST2520 UB12-TC strain, respectively, E. coli B1 ST2540 NI-4 clone and E. coli B2 ST95 NI-9 clone) (Fig. 3A; Table S1).

Variable cost of IncI plasmid in transconjugants E. coli NI from the pig’s gut

To evaluate the impact of the IncI pST3 plasmid on new bacterial host, we proposed to compare the MGRs of the UB12-TC strain, the M63-DN donor, and 11 cefotaxime-resistant and rifampicin-susceptible NI from the stools of the pigs, between day 4 and day 22. These isolates all harbored the IncI pST3 plasmids that have disseminated between different E. coli in the pigs’ microbiota after inoculation (Table 1). The MGRs of four NI were not significantly different from that of the donor strain M63-DN, while seven NI showed an MGR significantly lower than those of the M63-DN strain (P between 2.7 ×10⁻⁵ and 0.015) (Fig. 3B).

In order to determine the relative abundance of the newly formed transconjugants in the pig’s gut, we compared the titer of cefotaxime-resistant and rifampicin-susceptible E. coli clones to all cefotaxime-resistant E. coli population, on the days during which each NI was collected. We observed a variable relative abundance, ranging from 10% to 30% for NI-1, NI-2, NI-3, and NI-6, from 40% to 50% for NI-4, NI-9, NI-10, and NI-11, and more than 65% for NI-5, NI-7, and NI-8.

Transcriptomic analysis of UB12 strains

We compared the transcriptome of the ancestral UB12-TC transconjugant with the UB12-RC recipient strain to determine the impact of plasmid acquisition on host gene expression. We also compared the transcriptome of two UB12-TC^ive transconjugants, isolated between day 28 and day 30 in the pig’s gut, and of one UB12-TCΔpIncI transconjugant with the UB12-RC recipient and the ancestral UB12-TC strains to determine the impact of the IncI pST3 plasmid maintenance on the bacterial host in the pig’s gut. We combined the results from differential gene expression analysis of two independent biological replicates for each clone, except for UB12-TC.

We observed a low significant transcriptional response of the acquisition of plasmids into UB12-TC, with 10 significantly (> 2 log change) differentially expressed genes [adjusted P value (P _adj) <0.05] between UB12-RC and UB12-TC strains. Six genes were down-regulated and four genes were up-regulated in UB12-TC compared to UB12-RC. Functions affected by plasmids acquisition included genes involved in mobility and chemotaxis (down-regulation of flgB gene; up-regulation of motA and cheA genes), in stress response (down-regulation of cyuP and ydcI genes; up-regulation of gnsA gene) or genes encoding for phage proteins (down-regulation of three chromosomal genes; up-regulation of one IncY plasmid-gene) (Table 3).

TABLE 3.

List of differentially expressed genes from UB12-TC clones compared to UB12-RC recipient strain ^a

UB12-RC					UB12-TC		UB12-TCive.1		UB12-TCive.3		UB12-TCΔpIncI.1
Genome reference	Locus reference	Gene	Product	GO terms b	log2FC	P adj	log2FC	P adj	log2FC	P adj	log2FC	P adj
ECOL4764K	85852650	cheA	Chemotaxis protein CheA	Chemotaxis (NT)	2,44784	0,00238	1,66181	0,02643	1,61793	0,05792	3,39695	3,1E-17
ECOL4764K	85852652	tar	Methyl-accepting chemotaxis protein Tar	Chemotaxis (NT)	0,35905	0,99976	0,90038	0,56711	1,08194	0,12425	2,52615	1,8E-09
ECOL4764K	85852654	cheR	Chemotaxis protein methyltransferase	Chemotaxis (J)	1,31467	0,99976	0,53732	0,99727	0,74834	0,94242	1,94461	2,6E-05
ECOL4764K	85852655	cheB	Protein-glutamate methylesterase/protein glutamine deamidase	Chemotaxis (NT)	1,61018	0,86086	1,95395	0,00039	1,90874	0,0255	3,37828	4,9E-13
ECOL4764K	85852656	cheY	Chemotaxis protein CheY	Chemotaxis (KT)	−0,386	0,99976	0,84639	0,72816	0,84993	0,41767	1,44826	0,00534
ECOL4764K	85852657	cheZ	Chemotaxis protein CheZ	Chemotaxis (J)	0,14321	0,99976	0,47045	0,99727	0,53222	0,99983	1,46782	0,01351
ECOL4764K	85853147	trg	Methyl-accepting chemotaxis protein Trg	Chemotaxis (NT)	0,22622	0,99976	0,47636	0,99727	0,34401	0,99983	1,32684	0,01396
ECOL4764K	85852653	tap	Methyl-accepting chemotaxis protein Tap	Chemotaxis (NT)	1,64986	0,27944	1,79255	0,00465	1,92087	0,00359	3,55573	1,2E-20
ECOL4764K	85852566	yeeN	Putative transcriptional regulator YeeN	Function unknown (K)	1,56735	0,18152	1,87188	0,08396	1,94558	0,01514	1,2682	0,99982
ECOL4764K	85851949		Protein of unknown function	Function unknown (S)	1,18937	0,64644	0,95973	0,3662	1,05386	0,0478	0,54491	0,99982
ECOL4764K	85852023		Conserved protein of unknown function	Function unknown (S)	1,0608	0,72326	1,91186	0,12351	1,61571	0,03232	0,69775	0,99982
ECOL4764K	85852530	yeeX	DUF496 domain-containing protein YeeX	Function unknown (S)	1,48447	0,12054	1,32007	0,00832	1,55266	0,00204	1,26712	0,42012
ECOL4764K	85853785		SIR2_2 domain-containing protein	Function unknown (S)	1,78236	0,05217	2,05694	0,00026	2,08265	3,5E-05	1,6974	0,99982
ECOL4764K	85854633	yjcZ	Uncharacterized protein YjcZ	Function unknown (S)	1,01766	0,99976	0,54035	0,99727	0,85744	0,39624	2,44203	0,0003
ECOL4764K	85854701		YjbI protein	Function unknown (S)	−0,9134	0,99976	2,28377	0,03834	2,27538	0,05809	0,7817	0,99982
ECOL4764K	85854863	ybdZ	Enterobactin biosynthesis protein YbdZ	Function unknown (S)	−2,2921	0,99976	3,13395	0,02211	1,40956	0,22322	0,91143	0,99982
ECOL4764K	85851865	srlA	Sorbitol-specific PTS enzyme IIC2 component	Metabolism (G)	0,4513	0,99976	1,62365	0,99727	1,27625	0,03441	1,14754	0,81138
ECOL4764K	85852338	napD	NapA signal peptide-binding chaperone NapD	Metabolism (P)	0,82012	0,99976	2,2929	0,056	2,87501	0,00812	2,25275	0,06792
ECOL4764K	85853341	narK	Nitrate:nitrite antiporter NarK	Metabolism (P)	−0,3839	0,99976	1,057	0,97335	2,61826	0,04322	1,92704	0,78516
ECOL4764K	85854703	lamB	Maltose outer membrane channel/phage lambda receptor protein	Metabolism (M)	−0,332	0,99976	3,02916	0,03104	2,854	0,01182	1,67913	0,99982
ECOL4764K	85854704	malK	Maltose ABC transporter ATP binding subunit	Metabolism (P)	−0,9127	0,99976	2,91751	0,00246	2,87201	0,00369	1,26125	0,01394
ECOL4764K	85854705	malE	Maltose ABC transporter periplasmic binding protein	Metabolism (P)	−0,0777	0,99976	2,29064	0,05361	2,32236	0,03579	1,12427	0,06162
ECOL4764K	85854706	malF	Maltose ABC transporter membrane subunit MalF	Metabolism (G)	0,13807	0,99976	2,87309	0,00955	2,83601	0,01087	1,03171	0,41672
ECOL4764K	85853611	gnsA	Putative phosphatidylethanolamine synthesis regulator GnsA	Metabolism (S)	1,94991	0,01334	1,62445	0,56803	1,48141	0,4418	0,75222	0,99982
ECOL4764K	85852618	fliS	Flagellar biosynthesis protein FliS	Mobility (N)	−0,3836	0,99976	1,10601	0,47705	1,05056	0,31126	1,59581	0,00036
ECOL4764K	85852620	fliC	Flagellar filament structural protein	Mobility (N)	1,27948	0,36049	1,36333	0,00715	1,47471	0,00223	2,6014	6,3E-11
ECOL4764K	85852648	motA	Motility protein A	Mobility (N)	2,39223	0,01244	2,45117	2,1E-06	2,67929	5,7E-11	4,82804	5,9E-22
ECOL4764K	85852649	motB	Motility protein B	Mobility (N)	1,88766	0,12858	1,62233	0,0069	1,54818	0,02353	3,54867	6,12E-19
ECOL4764K	85853521	flgL	Flagellar hook-filament junction protein 2	Mobility (N)	0,27039	0,99976	0,61263	0,99727	0,5725	0,99983	1,70499	1,9E-05
ECOL4764K	85853522	flgK	Flagellar hook-filament junction protein 1	Mobility (N)	0,10274	0,99976	0,36238	0,99727	0,50076	0,99983	2,0174	0,02322
ECOL4764K	85853533	flgM	Anti-sigma factor for FliA (sigma (28))	Mobility (N)	0,34001	0,99976	0,69974	0,99727	0,58044	0,99983	1,97572	0,0005
ECOL4764K	85853534	flgN	Flagellar biosynthesis protein FlgN	Mobility (N)	−0,7733	0,99976	0,781	0,99727	0,90817	0,888	1,786	0,01351
ECOL4764K	85854634	crfC	Clamp-binding sister replication fork colocalization protein	Mobility/Biofilm (S)	1,70975	0,07724	1,95494	0,00099	2,21853	0,0002	4,27686	4,6E-07
ECOL4764K	85851287	pdeH	c-di-GMP phosphodiesterase PdeH	Signal transduction/ Biofilm (T)	1,46386	0,99976	1,66924	0,00446	1,71645	0,00094	3,34377	9,2E-09
ECOL4764K	85855159	aer	Aerotaxis sensor receptor, flavoprotein	Signal transduction (T)	0,77505	0,99976	0,55373	0,99727	0,57299	0,99983	2,01637	7,1E-08
ECOL4764K	85853407	ariR	Regulator of acid resistance, influenced by indole	Stress response (K)	−1,4935	0,99976	2,80049	0,04361	1,93625	0,08163	0,09459	0,99982
ECOL4764K	85854675	soxS	DNA-binding transcriptional dual regulator SoxS	Stress response (K)	0,87424	0,99976	2,21651	0,0026	2,98162	2E-08	0,28149	0,99982
ECOL4764K	85853493	bhsA	DUF1471 domain-containing multiple stress resistance outer membrane protein BhsA	Stress response (M)	1,59777	0,29059	2,42131	0,00119	2,25489	2,7E-06	0,96289	0,99982
ECOL4764K	85855319	yhcN	DUF1471 domain-containing stress-induced protein YhcN	Stress response (S)	1,87275	0,07207	3,39818	0,00145	2,98628	0,00046	1,27996	0,99982
ECOL4764K	85851256	cspA	Cold shock protein CspA	Transcription (K)	1,12139	0,61924	1,56845	0,00214	2,01672	8,8E-06	1,01535	0,99982
ECOL4764K	85853716	infA	Translation initiation factor IF-1	Translation (J)	1,31764	0,30382	1,39097	0,02694	1,1611	0,16529	0,66463	0,99982
p4764_IncY.1	71940636		Phage protein	Phage-related gene	3,14456	0,0015	2,18085	0,00086	2,52625	0,00024	0,96609	0,99982
ECOL4764K	85854613	ghoT	Toxin of the GhoTS toxin-antitoxin system	Virulence (S)	0,83692	0,99976	1,12523	0,57857	1,84114	0,01536	1,14237	0,99982
ECOL4764K	85853786		AAA family ATPase		1,1885	0,64644	1,39583	0,06825	1,42295	0,03606	0,93341	0,99982
ECOL4764K	85853146	ydcI	Putative DNA-binding transcriptional repressor YdcI	Transcription (K)	−2,6947	0,01855	−1,0757	0,99727	−1,2121	0,57142	−0,9326	0,99982
ECOL4764K	85852863	lpp	Murein lipoprotein	Stress response (M)	−1,7995	0,26242	−3,3785	0,00014	−2,1319	0,27055	−2,8267	0,00396
ECOL4764K	85854468	fimA	Type 1 fimbriae major subunit	Adherence, virulence (NU)	−0,1746	0,99976	−1,3709	0,02694	−1,2625	0,04858	−1,1741	0,05009
ECOL4764K	85853794		Conserved protein of unknown function	Function unknown (S)	−2,6511	0,99976	−3,0843	0,03081	−2,7848	0,03065	−1,0916	0,99982
ECOL4764K	85854469		Conserved protein of unknown function	Function unknown (S)	−0,884	0,99976	−1,593	0,02238	−1,3599	0,05618	−1,0845	0,99982
ECOL4764K	85855200	cyuP	Putative D/L-serine transporter	Metabolism (E)	−2,4647	0,00398	−0,4576	0,99727	−0,0101	0,99983	0,09587	0,99982
ECOL4764K	85855214	garP	Galactarate/glucarate/glycerate transporter GarP	Metabolism (P)	−3,286	0,99976	−3,2159	0,03322	−2,191	0,40415	−1,2117	0,99982
ECOL4764K	85853531	flgB	Flagellar basal-body rod protein FlgB	Mobility (N)	−2,4376	0,01135	0,29403	0,99727	−0,0528	0,99983	1,05304	0,99982
ECOL4764K	85853758		Phage tail assembly protein	Phage-related gene	−5,9563	0,19907	−2,0276	0,05361	−2,0384	0,01677	−0,6334	0,99982
ECOL4764K	85853759		Major tail tube protein	Phage-related gene	−3,1385	0,02476	−3,3181	7,1E-05	−3,3456	1,4E-05	−0,7553	0,99982
ECOL4764K	85853760		Phage tail protein	Phage-related gene	−3,3751	0,01825	−3,3118	0,0003	−3,1419	0,00019	−0,615	0,99982
ECOL4764K	85853767		Baseplate assembly protein J	Phage-related gene	−2,738	0,64644	−2,4369	0,01225	−2,0464	0,01131	−0,4411	0,99982
ECOL4764K	85853769		Baseplate assembly protein V	Phage-related gene	−5,2672	0,40761	−1,9515	0,03139	−1,7238	0,05644	−0,4465	0,99982
ECOL4764K	85853774		Putative regulatory protein	Phage-related gene	−4,5905	0,73682	−2,4347	0,08358	−2,9586	0,00028	−0,7132	0,99982
ECOL4764K	85853775		Putative membrane protein	Phage-related gene	−5,7573	0,28603	−3,6302	0,00097	−3,3555	0,00137	−1,1559	0,99982
ECOL4764K	85853776		Prophage lysozyme ; Phage lysin	Phage-related gene	−4,9729	0,64535	−3,7054	0,00406	−3,1939	0,00726	−0,5838	0,99982
ECOL4764K	85853779		Phage head completion-stabilization protein	Phage-related gene	−5,7646	0,27944	−3,3571	0,00598	−3,2539	0,00252	−0,8388	0,99982
ECOL4764K	85853780		Phage terminase, endonuclease subunit	Phage-related gene	−6,0071	0,22056	−4,1881	0,00013	−4,1537	4E-05	−0,8143	0,99982
ECOL4764K	85853781		Phage major capsid protein, P2 family	Phage related gene	−4,9297	0,00621	−4,0468	4,2E-05	−4,3081	7,3E-07	−0,8018	0,99982
ECOL4764K	85853782		Phage capsid scaffolding protein	Phage-related gene	−3,2044	0,09625	−3,8346	6,9E-05	−3,9739	2E-05	−0,7085	0,99982

^a Boxes in gray color indicate genes that were statistically significantly different compared to UB12-RC (positive value = up-regulated; negative value = down-regulated) with p _adj <0.05 (dark gray) and with p _adj <0.1 (light gray); log2FC indicates the log₂ (fold-change), p _adj indicates the adjusted P-value

^b Gene Ontology terms are according to the BioCys Genome Database collection and eggNOG-mapper v2.

No significant change in chromosomal and plasmid gene expression was observed in the UB12-TC^ive clones, compared to UB12-TC before carriage in the pig’s gut.

However, we detected a transcriptional response between the UB12-RC and UB12-TC^ive strains, with a total number of significantly (> 2 log change) differentially expressed genes (P _adj <0.05) ranging from 37 to 39. Among these differentially expressed genes observed between UB12-RC and both UB12-TC^ive clones, most of down-regulated chromosomal genes encoded for phage proteins (12 genes). Other down-regulated genes were fimA gene, encoding for type 1 fimbriae major subunit, and two genes involved in metabolism (lpp and garP genes, significantly detected down-regulated only for one of UB12-TC^ive clones) (Table 3). In one or both of UB12-TC^ive clones, we detected up-regulated genes involved in various stress response, such as ariR (oxidative stress), bhsA (multiple stress resistance), yhcN (hydrogenmean of absorbance peroxide stress), cspA (temperature response), and soxS (superoxide stress master regulator), and genes encoding for transporters in various metabolic pathways (srlA, for sorbitol; narK or napD, for nitrate/nitrite; and malEFK, for maltose). Other functional classes of genes detected as significantly up-regulated were those involved in biofilm (crfC and pdeH genes), in mobility (motAB and fliC genes), and in chemotaxis (cheB and tap genes) (Table 3).

All genes involved in mobility and chemotaxis, differentially expressed between UB12-TC^ive and UB12-RC strains, were also up-regulated in the UB12-TCΔpIncI clone. Moreover, other genes included in the chemotaxis system (tar, cheZYR, and tgr genes) and the flagella systems (fliS, flgN, flgKL, and flgM genes) were up-regulated between UB12-TCΔpIncI and UB12-RC strains. Interestingly, fliA, the other major regulator of the flagella systems, appeared significantly differentially expressed (down-regulation) between UB12-TC and UB12-TCΔpIncI clones.

Using the replicase repA gene for comparison, we showed that the transcription of IncI plasmid genes was not different between UB12-TC and UB12-TC^ive clones (Table S2). We observed that most of the IncI plasmid genes (111 genes, 89.5%) were expressed at low level or not expressed compared to repA transcription. Only 13 of predicted ORFs (10%) were transcribed at level between 4- and 35-fold higher than repA. The most expressed plasmidic gene was annotated as a conserved protein of unknown function, located just upstream of oriT and the nikA gene (encoding for a Ribbon-helix-helix relaxosome protein), which was transcribed at level 5-fold higher than repA. Two other loci in plasmid core regions were transcribed at levels between 5- and 16-fold higher than repA, the excAB genes and an ORF located just upstream, and the parBM genes. Two accessory regions of the IncI plasmid encoding resistance genes were also transcribed at high level, the ISEcp-1-bla _CTX-M-1 cassette and three genes (glmM, sul2, and drfA) located on an ISCR2 mobile genetic element (Table S2). We observed that the IncY plasmid genes were also expressed at low level. Only 10% (11 genes) of the coding sequences were transcribed at higher level than repA, and all annotated as conserved proteins of unknown function (Table S2).

Swimming motility of strains

To confirm the transcriptomic data on genes involved in mobility, we tested the swimming ability of all the strains used in our study in semi-solid LB agar tubes. We observed that the swimming ability decreased in UB12-TC strains compared to UB12-RC and M63-DN strains. While the ancestral UB12-TC strain presented a comparable swimming motility phenotype to all UB12-TC^ive clones, UB12-TCΔpIncI clones presented an increase in the ability to swim in semi-solid LB agar tubes. Of note, we observed various phenotypes, from nonmotile to motile, for all 11 NI pig’s clones, which all harbored IncI plasmid (Table 1).

Biofilm formation of UB12 strains

To investigate the effect of the IncI plasmid on the biofilm formation, we compared the biofilm-forming ability of M63-DN, UB12-RC, UB12-TC strains and UB12-TCΔpIncI strains. After an incubation time of 48 h, biofilm formation was assessed by staining the attached cells with crystal violet (CV). All strains formed poor biofilms with an OD₅₇₀ between 0.079 +/− 0.055 [means +/− standard deviation (M +/− SD)] for M63-DN, and 0.114 +/− 0.66 (M +/− SD) for UB12-RC (Table 1). Furthermore, no significant differences were observed between plasmid-carrying and plasmid-free isolates (P = 0.91).

DISCUSSION

Multiresistance plasmids are widely distributed across bacteria and are one of the main drivers of antibiotic resistance spread. The success of a plasmid-borne resistance is constrained by the fitness cost that the multiresistance plasmid imposes on the bacterial cell. It is not clear whether epidemic bla _CTX-M-1-IncI1 plasmids, one of the most common plasmid type in ESBL producers of animal origin (2, 8), conferred a fitness cost on the host bacterium. Previous studies have shown that the bla _CTX-M-1-IncI1 plasmid imposes no or negligible fitness costs on its E. coli host in vitro (28) and persists without antimicrobial usage (29). On the other hand, using in vitro competition experiments, Freire Martín et al. showed that the IncI1 plasmid-cured isolate outcompeted the original plasmid-carrying isolate, or plasmid-complemented strains, suggesting a fitness burden was imposed by carriage of this plasmid (30). These studies were performed using in vitro models and/or laboratory-adapted strains with limited clinical relevance. Here, we investigated the cost of a bla _CTX-M-1-IncI1/ST3 epidemic plasmid in a commensal E. coli animal strain, before and after oral inoculation of pigs. In our model, bla _CTX-M-1-IncI1/ST3 plasmid imposed no measurable burden on the recipient strain after conjugation and we observed no modification of the bacterial growth rate of the bla _CTX-M-1-IncI1 plasmid-carrying strain after the longitudinal carriage in the pig’s gut. However, a significant difference of bacterial growth rate was found between the bla _CTX-M-1-IncI1/ST3 plasmid-carrying strain and the transconjugants that had spontaneously lost the IncI1 plasmid during in vivo carriage, suggesting a small fitness burden exists. The loss of IncI plasmid is a rare event (10), as expected because of the presence of addiction systems on these large conjugative multiresistance plasmids. In our model, a ColE1-like plasmid (ColRNAI plasmid) has disseminated, together with bla _CTX-M-1-IncI1/ST3, from the donor strain to our recipient strain. We could not independently investigate the fitness effect of IncI and ColRNAI on the recipient strain. In the literature, few studies on the impact of ColE1-like plasmids on bacterial recipient strains have demonstrated that their acquisition imposes a fitness cost, which could be alleviated during in vitro experimental evolution (31, 32). We have not observed this in our experiments. Recently, some authors have also proposed that ColE1 plasmids contributed, in association with chromosomal and plasmid mutations, to the fitness of Salmonella spp. in a specific environmental niche (33). In contrast to this experiment, we detected a very small number of chromosomal mutations, which were not convergent, and no plasmid mutation among the transconjugants. Although we cannot exclude that the ColRNAI plasmid influences the fitness of the transconjugants, its role seems negligible in our model. As a whole, our observations could suggest that bla _CTX-M-1-IncI1 epidemic plasmids do not induce a sufficient fitness cost on its host bacterium to drive their extinction in gut microbiota, in the absence of antibiotics.

After ESBL-producing E. coli inoculation, bla _CTX-M-1-IncI1/ST3 plasmid disseminated rapidly to natural isolates in the pig’s gut, highlighting the usually very efficient conjugation system of IncI plasmids (8, 34). As a reference to the bacterial growth rate of the bla _CTX-M-1-IncI donor strain, the comparative growth rate of the 11 newly formed transconjugants in the pig’s gut appeared extremely variable, which can be related to the variable fitness of these IncI plasmid-carrying new spontaneous transconjugants. Although it is not possible to determine the competitiveness of the host cells before IncI plasmid acquisition due to the experimental design of our study, this suggests that the fitness cost could not act as a strong barrier to transmission of the bla _CTX-M-1-IncI1 epidemic plasmid, among natural populations of E. coli in a relevant environmental niche such as the gut. Interestingly, the relative abundance of some of the newly formed transconjugants was greater than 50%, suggesting that these clones represented dominant ESC-resistant E. coli strains in the pig’s gut at this time.

Previous studies have reported in vivo transfers of multidrug-resistant plasmids despite competition from fecal flora, and a high risk of rapid increase of transconjugants becoming the dominant population under antibiotic treatment (34 – 36). Some studies have shown that antibiotics exposure increased conjugative transfer of bla _CTX-M-1-IncI plasmids (37) and it has been suggested that the transfer rate of the IncI plasmid may be underestimated in in vitro study models (38). The high predominance and worldwide spread of some IncI1 plasmids, especially among animal ESBL-producing E. coli, could be explained by a low fitness cost imposed on the new recipient strain, combined with a high conjugation efficiency and potentiated by a high exposure of animals to antibiotics in agricultural settings.

A complex plasmid-chromosome cross-talk is involved in plasmid domestication into their host cells and often appears to be dependent of the plasmid–host pairing studied. Among these plasmid–host adaptation effects associated with a potential fitness benefit, only changes in chromosomal gene expression have been observed in some plasmid–host evolutions (39). These modifications in the expression of chromosomal genes for the bacterium adaptation to the acquisition of the plasmid appeared highly variable both in the degree of alteration and in the range of cellular functions that are affected (39 – 44). However, chromosome-encoded functions most commonly targeted by plasmid carriage include respiration, signaling, energy production, various metabolism pathways, and virulence factors, especially adhesion-related functions involved in biofilm formation and mobility (45, 46). To analyze the impact of the bla _CTX-M-1-IncI1/ST3 plasmid on the commensal E. coli host, we performed WGS and RNA-sequencing of the ancestral strains on the one hand, and UB12 isolates collected from pigs’ feces on the other hand. In our study, a very small number of SNPs in chromosomal and plasmid coding sequences were detected in the UB12 isolates, without clear convergence. In addition, RNA-sequencing revealed that only the expression of a small set of chromosomal genes was altered in the host cell after acquisition of the IncI ST3 plasmid. There are chromosomal genes encoding prophage proteins and genes involved in different processes such as metabolism, stress response, and cell mobility, which is consistent with observations from previous studies with others plasmids (39, 41, 47, 48). We showed that the transcriptional expression of chemotaxis and flagellar genes change between IncI plasmid-carrying and plasmid-free UB12 isolates collected after oral inoculation of pigs, in accordance with the modification in the swimming capacities of these clones. The presence of conjugative plasmids has various effects on mobility of the recipient strains (18, 49). Many studies have also reported their role in biofilm development with various and opposite effects, which could depend strongly on the host as well as on the plasmid (50). We tested the ability of our strains to form biofilms, but all UB12 isolates showed low biofilm formation, under our conditions, and we did not observe any significant difference between the strains. A complex interplay exists between conjugation, mobility, and biofilm. Some plasmid-borne regulators interact with chromosomal genes responsible for flagellar biosynthesis, chemotaxis and mobility and interferes with the plasmid transfer dynamics (51 – 53).

None of the mutations observed in our data were associated with change in gene expression, suggesting that we observed a physiological response. It is interesting to note that chromosomal genes implicated in cell mobility are also involved in the bacterial response to environmental changes especially in the highly dynamic and compartmentalized gut environment (54). In our model, plasmid–host adaptation could be the result of both the specific plasmid–host pairing studied and adaptation to environmental changes in the pig’s gut.

Overall, we observed that the bla _CTX-M-1-IncI1/ST3 plasmid has only a limited effect on the host cell, targeting cellular functions classically known to be affected in the host bacterium after plasmid acquisition.

We observed that most of the bla _CTX-M-1-IncI1/ST3 plasmid genes were expressed at low level or not expressed in host cell. Our observation indicates that the IncI plasmid transcription is highly regulated, which is consistent with findings from other studies reporting transcriptional profiles of other large conjugative plasmids (43). Plasmid transcriptional control is common. For conjugative plasmids, a regulatory network based on strong plasmids promoters, which are repressed in negative feedback loops, has been described and are transiently modulated during the lifestyle of conjugative plasmids (55, 56). Plasmids induce a fitness deficit on their hosts, distinguishing between the long-term fitness cost and the plasmid acquisition cost (57). Therefore, plasmid dissemination is driven by two intrinsically conflicting processes, vertical spread by multiplication of plasmid-carrying host cells and horizontal spread by invading new cells. Although many previous studies have focused on the phenomenon of compensatory evolution of costly plasmids in recipient strains contributing to the survival and spread of these plasmids (13, 19 – 23), recent studies demonstrated that the plasmid acquisition cost and the short-term effects within the cell could be critical to explain ecological plasmid success among bacterial population (58, 59). Moreover, in a recent work, Ahmad et al. demonstrated that maintenance of some degree of cost is beneficial for the selection of some new transconjugants in the bacterial population. They observed that intermediate-cost plasmids were selected compared to their higher- and lower-cost counterparts (58). The success of IncI plasmids could be dependent of a complex regulatory network that conciliates a perfect balance between the expression of plasmid functions and the burden imposed on the host cell, maintaining an intermediate cost of these plasmids.

Recent studies have reported the involvement of plasmid genes encoding for nucleoid-associated proteins (NAPs) homologs, such as H-NS-like proteins, in transcriptional regulation networks between plasmids and host chromosomes (46, 60 – 62). Plasmids encoding H-NS-like “stealth” protein reduce their fitness cost by silencing transcriptional activities of their own genes, as well as some chromosomal genes, and therefore minimize the impact of plasmid introduction in the bacterium. Large conjugative plasmids frequently carry NAP gene homologs but, using in silico analysis, similar genes were not yet identified in plasmids of the IncI incompatibility group (63). However, a large number of IncI backbone genes are yet annotated as hypothetical proteins. As a whole, future studies could be interesting to characterize new regulators still unknown among these genes, and to better understand the cross-talk between IncI plasmids and the host chromosome.

Our study has further limitations. First, we performed the experiments in a unique commensal E. coli host strain and the genetic background of strains plays an important role in plasmid–host adaptation. Second, due to the design of our study, we used in vitro plating of the bacteria and therefore we could not directly observe the E. coli populations of interest in the pig’s gut. This approach might have introduced a bias as a consequence of short adaptation of the IncI plasmid in the new recipient strain. Third, given that there are relatively large differences between the core genomes of NI isolates, it would have been appropriate to compare each natural isolate with its IncI plasmid-free counterpart, to assess the fitness cost of IncI1 plasmid carriage according to the host strain’s background. Unfortunately, we were unable to obtain IncI plasmid-free NI clones, except for one isolate, that had lost all its plasmid content, i.e., IncI and IncF (data not shown). Nevertheless, our study is the first to provide direct evidence of in vivo plasmid transfer in the pig’s gut and to confirm that bla _CTX-M-1-IncI1/ST3 plasmid fitness cost on its bacterial hosts is negligible in vivo, in a relevant environmental niche such as the gut.

MATERIALS AND METHODS

Bacterial strains and plasmids

For this study, a total of 24 strains were selected from a pig strain collection of ANSES laboratory (Ploufragan, France). The E. coli transconjugant UB12-TC was obtained by conjugation between a randomly chosen pan-susceptible pig commensal E. coli isolate, made resistant to 250 mg/L rifampicin (recipient strain UB12-RC, B1, O100:H40) and an ESBL-resistant E. coli pig isolate (donor strain M63-DN, B1, O147:H7). UB12-RC and M63-DN belong to ST2520 and ST2521, respectively, according to the Warwick MLST scheme.

Thus, E. coli UB12-TC is resistant to rifampicin and harbors a IncI1 pST3 plasmid encoding bla _CTX-M-1, sul2, dfrA17, and aadA5 genes. Seven cefotaxime- and rifampicin-resistant isolates (named UB12-TC^ive), three cefotaxime-susceptible and rifampicin-resistant isolates (named UB12-TCΔpIncI), and eleven cefotaxime-resistant and rifampicin-susceptible isolates (named NI-1 to NI-11) were chosen among all strains collected during a series of experiments conducted to evaluate the carriage of ESBL-producing E coli in the pig’s gut (11, 12). These isolates have been collected from day 4 to day 45 after pig inoculation.

Animals and experimental design

As previously described in Mourand et al. (10, 12), the experiments were conducted in the ANSES, Ploufragan, France, animal facilities. The trials were conducted with 6- to 8-week-old specific pathogen-free and initially ESC-resistant E. coli-free Large White piglets. The animals were randomized before the experiment. They did not receive any antibiotic treatment prior to the trial and were given the same nonsupplemented feed. Each room contained eight pigs placed in two pens of four animals each. For both trials, UB12-TC was inoculated on the first day (day 0): each pig was orally given a 10 mL suspension prepared from strain (between 2.5 × 10⁶ to 7.8 × 10⁹ CFU/mL) cultivated on MH agar containing cefotaxime (2 mg/L). Individual fecal samples were collected from animals on day 0 before inoculation, and regularly, usually once or twice a week, between day 1 and day 45. The fecal samples were stored at −70°C until analysis.

The numbers of cefotaxime-resistant E. coli (including UB12-TC and NI) in the individual fecal samples were estimated by spreading 100 µL of 10-fold dilutions on MacConkey agar plates (containing respectively 2 mg/L cefotaxime), in triplicate. The titers of total E. coli population were determined by plating serial dilutions on MacConkey agar plates without antibiotic. After incubation at 37°C, the colonies obtained on supplemented or nonsupplemented MacConkey agar plates were enumerated, and the titers were calculated for each pig per day. The relative abundance of cefotaxime- and rifampicin-susceptible (NI clones) was calculated based on the ability of 2–5 isolates obtained from each pig on each day on cefotaxime-supplemented media to grow on rifampicin-supplemented media, and expressed as a percentage. For both trials, strains of interest were stored at −70°C in BHI + glycerol, at the end of the quantification in the fecal samples.

Antimicrobial susceptibility testing

Antimicrobial susceptibility testing by the disk diffusion method (ampicillin, ticarcillin, piperacillin, amoxicillin-clavulanic acid, ticarcillin-clavulanic acid, piperacillin-tazobactam, cefotaxime, ceftazidime, aztreonam, cefoxitin, cefepime, ertapenem and imipenem, association of sulfonamides + trimethoprim, tetracycline, streptomycin, nalidixic acid, ciprofloxacin, and rifampicin) was performed according to European Committee on Antimicrobial Susceptibility Testing (EUCAST) guidelines (http://www.eucast.org).

Motility assay

Donor strain M63-DN, recipient strain UB12-RC, ancestral UB12-TC, all UB12-TC^ive and UB12-TCΔpIncI clones were grown overnight in LB medium, washed, suspended in isotonic water to a concentration of 1 × 10² cells/mL and used to inoculate semi-solid LB agar tubes containing 0.35% agar, incubated for 48 h at 37°C. Experiments were repeated three times. For data analysis, three categories were considered (motile, partially motile, and nonmotile).

Biofilm formation assay

Biofilm formation assay was performed using a 96-well flat-bottom microtiter plate (Corning, ME, USA) by staining polystyrene-attached cells with crystal violet (CV). From a glycerol stock, the UB12 strains were plated on LB agar and grown overnight at 37°C. Randomly selected colonies were grown overnight in 10 mL of M9 minimum media (supplemented with 0.1% glucose, MgSO₄ 1M, and CaCl₂ 100 mM) at 37°C under constant stirring at 200 rpm. Next, 100 µL of each culture was transferred into 10 mL of fresh M9 minimum media (supplemented as above) and 200 µL of each culture were transferred in the plate. Then, the biofilm assay was performed over a 48 h period at 37°C without stirring, which allowed biofilm formation. The bacterial suspensions were removed and the wells were washed three times with 200 µL of PBS. Microtiter plates were heated at 50°C for 1 h to fix the biofilm and then surface-attached cells were incubated with 100 µL of 0.1% CV and air dried for 15 min. After five washes with distilled water and 2 h of drying at room temperature, the dye was eluted with 100 µL of 50% ethanol. Absorbance measurements (DO 570/595 nm) obtained with a microplate reader (Infinite F200-Pro, Tecan) were normalized by values obtained from wells containing sterile M9 minimum media.

The biofilm experiments were performed in three biological replicates with three technical replicates each.

In vitro evolution experiments

Five randomly selected colonies of UB12-RC and UB12-TC were inoculated into 5 mL of LB and grown under constant shaking at 200 rpm. After overnight incubation at 37°C, 5 µL of each culture was inoculated into five tubes containing 5 mL of fresh LB. At this step, the 10 cultures were considered the ancestors of each lineage (day 0 of the evolution assay). Next, the 10 cultures were propagated by daily serial transfer, with a 1/1,000 dilution (5 µL of culture grown overnight diluted into 5 mL of fresh LB without antibiotic), for 30 days (corresponding to 300 generations in total) and incubated overnight at 37°C under constant stirring at 200 rpm. Finally, after 30 days, 10 independent lineages were obtained (denoted “ite”). At day 0 and every 5 or 6 days of the evolution assay, the culture was isolated on LB agar for purity, and antibiotic susceptibility was tested by disk diffusion. A culture sample from each lineage was collected and stored at −80°C with glycerol for future analysis.

Individual fitness assays

Fitness, defined here as the maximum growth rate (MGR), was determined from growth curves using a high-precision technique developed in-house as described previously (18). For each strain, the fitness assay was repeated six times. The measurements were collected using a Python script developed in our laboratory, and the MGR was obtained by the lm function in R (http://www.R-project.org/).

Whole-genome sequencing

WGS was performed using Illumina technology. Briefly, DNA was prepared for paired-end sequencing on the Illumina MiniSeq (Integragen, MA, USA) for all strains. DNA samples were extracted using the genomic DNA NucleoMag tissue kit from Macherey-Nagel (Hoerdt, France). DNA libraries were prepared and indexed, and tagmentation was processed using the Nextera DNA Flex library Prep kit (Illumina, USA). The sample concentrations were normalized in equimolar concentration at 1 nM. The pooled libraries were sequenced to a read length of 2 by 150 bp with Illumina MiniSeq High output reagent Kit. The genomes were sequenced at an average depth of 50X.

DNA sequence analysis and UB12-TC genome comparison

Illumina genome sequence reads were assessed for quality using FastQC v0.11.8 and subsequently trimmed with a cutoff phred score of 30 using TrimGalore v0.4.5. All genomes were de novo assembled using SPAdes v.3.13.1 software and analyzed with an in-house bioinformatics pipeline (https://github.com/iame-researchCenter/Petanc) adapted from Bourrel et al. (64). Briefly, genome typing was performed including MLST determination, phylogrouping, and serotyping. The pipeline was also used to get information about plasmid replicons, resistance and virulence genes, the prediction of a plasmid sequence or a chromosome sequence for each contig using PlaScope (65).

For each sample, sequenced reads were mapped to a reference sequence to analyze the mutations and rearrangements. For reference plasmid sequences, we used the plasmid sequences extracted from GenBank (http://www.ncbi.nlm.nih.gov/GenBank/) (ColRNAI, GenBank accession number J01566; IncY, GenBank accession number MH42254; Col156, GenBank accession number EU090225 and IncI, GenBank accession number KJ484635) and de novo assembled contigs as predicted to plasmid genome of UB12-TC strain with PlaScope, included in our pipeline. For reference chromosomal genome, first, we calculated the mash distance between UB12-TC genome sequence and E. coli genomes from RefSeq database using Mash v.2.0 and we selected the closest E. coli strain assembly ASM190094v1. Then, all predicted chromosomal contigs of the UB12-TC genome were rearranged as scaffold with RAGOUT v2.3 (https://github.com/fenderglass/Ragout) (66) and the reference genome (E. coli assembly ASM190094v1). All scaffolds of the UB12-TC strain obtained were annotated on the MicroScope platform (67). Identification of the point mutations and rearrangements between UB12-TC and UB12-TC^ive were performed using the open computational pipeline Breseq v0.27.1 (26). All SNPs identified in the ancestor clones that were sequenced were subtracted from the sequences of the evolved clones.

Transcriptome sequencing

RNA sequencing was performed on the recipient strain UB12-RC, the ancestral transconjugant UB12-TC, two evolved UB12-TC transconjugants (UB12-TC^ive, detected in pig’ stools at day 28 and day 31) and one UB12-TC transconjugant that had lost the plasmid IncI (UB12-TCΔpIncI, detected in pig’ stools at day 20). For the generation of transcriptomic data, selected bacteria were subcultured a total of three times. For RNA extraction, a total of 10 cultures, representing two biological replicates per strain, were independently grown. For each strain, 6 µL of an overnight culture was added into 6 mL of fresh LB and was incubated at 37˚C under constant stirring at 200 rpm until reaching an OD600 of approximately 0.6. Three milliliters of each culture were taken, representing two technical replicates, and mixed with RNA protect (Qiagen, Courtaboeuf, France) according to the manufacturer’s recommendations. Cells were immediately pelleted by a 10 min centrifugation at 5,000 x g. RNA was prepared using the Quick-RNA Fungal/Bacterial MiniPrep kit (Ozyme, ZymoResearch, USA). DNA was digested using Turbo DNAse (Thermofisher, UK) based on the manufacturer’s protocol. Verification of complete removal of any contaminating DNA was performed by PCR amplification of recA housekeeping gene. The final RNA solution was quantified with the Qubit RNA High Sensitivity assay kit (Thermofisher Scientific, UK) and RNA quality was evaluated using the Agilent 2100 Bioanalyzer RNA 6000Nano Kit (Agilent Technologies, Santa Clara, CA, USA). The RNA was immediately stored at −80°C.

For each replicate, 8 µL of RNA was treated to remove 16S and 23S rRNAs and cDNA libraries were prepared using the Zymo-Seq RiboFree Total RNA Library Kit (Ozyme, ZymoResearch, USA). The cDNA quantity was determined with the Qubit dsDNA Broad Range assay kit. Paired-end sequencing of the pooled libraries was performed with an Illumina MiniSeq Mid Output reagent kit and an Illumina NextSeq 500/550 High Output Kit v2.5 (Illumina, USA), configured to 2 × 150 bp cycles, yielding approximately 8 and 400 million reads, respectively.

Transcriptome analysis

To assure high sequence quality, the Illumina cDNA reads in FASTQ format were trimmed to remove all sequencing adapters using TrimGalore v0.4.5., according to the manufacturer’s recommendations. Then, cDNA reads were trimmed again, using TrimGalore v.0.4.5., with a cutoff phred of 30, and the reads shorter than 70 bp were discarded. A final quality check was done using FastQC v0.11.8. Reads were then mapped to the UB12-TC reference genome sequence (see above) using BWA alignment software and BWA-MEM algorithm v.0.7.17 (68). Mapping statistics were verified using BAMStat v.1.25. and SAMtools idxstat v.1.9 (69). Reads counts were performed using featureCounts v.2.0.3 (http://subread.sourceforge.net) (70). An R package, DESseq2 v.1.34. was used to normalize and to identify differentially expressed genes between each sample (http://www.bioconductor.org/packages/release/bioc/html/DESeq2.html) (71). A cutoff of q value of < 0.05 and a fold change of >2 were used to measure statistical significance.

Statistics

For statistical data analysis from the fitness assays, we used an analysis of variance (ANOVA) and unpaired Student’s test using R software. For statistical data analysis from the biofilm assays, we used ANOVA using GraphPad Prism software 9.0. P values of < 0.05 were considered significant.

ETHICS APPROVAL

The experiments were performed in accordance with French animal welfare regulations, and the protocol was approved by the ANSES/ENVA/UPEC ethical committee [ComEth authorizations 12-032 and 15-050 (no. APAFIS: 2015061813246553)]. The experiments were conducted in the ANSES, Ploufragan, France, animal facilities.

DATA AVAILABILITY

All reads have been deposited at the European Nucleotide Archive (project accession number PRJEB58070, ID samples from ERS14303872 to ERS14303895).

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/aac.00111-23.

Tables S1 and S2. aac.00111-23-s0001.xlsx.

Number of SNPs detected between UB12 clones, and new in vivo transconjugants; relative expression of plasmid genes.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.