Peptidoglycan Sensing Prevents Quiescence and Promotes Quorum-Independent Colony Growth of Uropathogenic Escherichia coli

Uropathogenic Escherichia coli (UPEC) is the leading cause of urinary tract infections (UTIs). During pathogenesis, UPEC cells adhere to and infiltrate bladder epithelial cells, where they may form intracellular bacterial communities (IBCs) or enter a nongrowing or slowly growing quiescent state. Here, we show in vitro that UPEC strains at low population density enter a reversible, quiescent state by halting division. Quiescent cells resume proliferation in response to sensing a quorum and detecting external signals, or cues, including peptidoglycan tetra- and pentapeptides.

ABSTRACT

Uropathogenic Escherichia coli (UPEC) is the leading cause of human urinary tract infections (UTIs), and many patients experience recurrent infection after successful antibiotic treatment. The source of recurrent infections may be persistent bacterial reservoirs in vivo that are in a quiescent state and thus are not susceptible to antibiotics. Here, we show that multiple UPEC strains require a quorum to proliferate in vitro with glucose as the carbon source. At low cell density, the bacteria remain viable but enter a quiescent, nonproliferative state. Of the clinical UPEC isolates tested to date, 35% (51/145) enter this quiescent state, including isolates from the recently emerged, multidrug-resistant pandemic lineage ST131 (i.e., strain JJ1886) and isolates from the classic endemic lineage ST73 (i.e., strain CFT073). Moreover, quorum-dependent UPEC quiescence is prevented and reversed by small-molecule proliferants that stimulate colony formation. These proliferation cues include d-amino acid-containing peptidoglycan (PG) tetra- and pentapeptides, as well as high local concentrations of l-lysine and l-methionine. Peptidoglycan fragments originate from the peptidoglycan layer that supports the bacterial cell wall but are released as bacteria grow. These fragments are detected by a variety of organisms, including human cells, other diverse bacteria, and, as we show here for the first time, UPEC. Together, these results show that for UPEC, (i) sensing of PG stem peptide and uptake of l-lysine modulate the quorum-regulated decision to proliferate and (ii) quiescence can be prevented by both intra- and interspecies PG peptide signaling.

IMPORTANCE Uropathogenic Escherichia coli (UPEC) is the leading cause of urinary tract infections (UTIs). During pathogenesis, UPEC cells adhere to and infiltrate bladder epithelial cells, where they may form intracellular bacterial communities (IBCs) or enter a nongrowing or slowly growing quiescent state. Here, we show in vitro that UPEC strains at low population density enter a reversible, quiescent state by halting division. Quiescent cells resume proliferation in response to sensing a quorum and detecting external signals, or cues, including peptidoglycan tetra- and pentapeptides.

INTRODUCTION

At least 80% of urinary tract infections (UTIs) are caused by uropathogenic Escherichia coli (UPEC) strains, and 27% of patients with a UTI experience recurrence within 12 months after successful antibiotic treatment (1–3). UPEC appears to enter a quiescent, nongrowing state within urothelial transitional cells in the bladder wall that may enable E. coli, long after antibiotic treatment has been terminated, to resume division within bladder urine after apoptotic release from those cells (4, 5). In this study, we show that various UPEC strains communicate by releasing and detecting external signals, or cues (6), which act as proliferants to regulate each bacterium’s decision to proliferate.

We recently reported the discovery of a quiescent state of E. coli that is common among UPEC strains (7). E. coli CFT073, a member of the major UPEC lineage sequence type 73 (ST73), and approximately 80% of ST73 strains tested proliferate on glucose minimal agar plates seeded with 10⁶ or more CFU but become quiescent on glucose minimal agar when plated at a cell density of less than 10⁶ CFU (7). Moreover, 23% of randomly selected UPEC strains of diverse STs isolated from community-acquired UTIs also become quiescent on glucose minimal plates seeded with less than 10⁶ CFU (7). Remarkably, these results indicate that growth, as measured by colony formation, of many UPEC strains on glucose minimal agar is quorum dependent (7). Alternative sole carbon sources, including acetate, arabinose, and N-acetylglucosamine (NAG), also lead to cells’ entering quiescence at low plating density, whereas glycerol, ribose, and xylose support proliferation (7). We also reported that mini-Tn5 transposon insertions in five central carbon metabolism genes (gdhA, gnd, pykF, sdhA, and zwf) prevent quiescence, suggesting that metabolic function must remain intact during quiescence even though the cells are nonproliferative. CFT073 quiescence is also prevented when cells are placed near a single colony of actively growing CFT073 or E. coli MG1655, a K-12 laboratory strain, suggesting that the actively growing cells release one or more molecules that act as a proliferant (7). Indeed, we showed that l-lysine is a proliferant for quiescent CFT073 cells, but only in combination with either l-methionine or l-tyrosine, yet CFT073 is not auxotrophic for lysine, methionine, or tyrosine, and the remaining 17 amino acids play no role in preventing CFT073 quiescence (7). Quorum-dependent proliferation suggests the ability to sense one or more cell-to-cell proliferants. In the case of CFT073, the proliferants include l-lysine and l-methionine, and detection of the proliferant underlies the decision by each individual cell to either grow and divide or to enter the quiescent state. This suggests that the transition between the proliferative and quiescent states is controlled by a quorum-sensing system (6, 8).

Here, we report that quiescent CFT073 cells are filamentous and that l-lysine import into quiescent CFT073 stimulates proliferation, suggesting that l-lysine synthesis is inhibited during quiescence. We show that in two pathogenic UPEC strains, CFT073 (a prototypic ST73 UPEC strain) and JJ1886 (a pandemic ST131 strain that displays increased sensitivity to proliferants), peptidoglycan (PG) stem peptides prevent quiescence. PG is a component of the bacterial cell wall that contains polymers of alternating linked sugars, NAG and N-acetylmuramic acid (NAM), with attached stem peptides containing d-amino acids that cross-link. PG fragments have been reported to regulate development of Bacillus subtilis and Mycobacterium tuberculosis (9–11). PG fragments are also normally secreted by growing E. coli (12), and here we show for the first time that they act as quorum-proliferant molecules in UPEC. In this study, we demonstrate that population sensing and metabolite availability modulate the quiescent state of UPEC, and we define an in vitro model system for UPEC quiescence for further study of UPEC physiology.

RESULTS

Colony formation of UPEC on glucose minimal agar plates is quorum dependent, and quiescence is reversed by secreted proliferants.

UPEC strain CFT073 enters a quiescent state when plated at low cell density (<10⁶ cells per plate) on glucose minimal agar yet grows robustly when plated at high cell density (>10⁶ cells per plate), demonstrating that growth, as measured by colony formation, of CFT073 on glucose minimal agar plates is quorum dependent (Fig. 1A). As reported previously, quiescent CFT073 cells can be resuscitated by challenge with specific external stimuli, called proliferants, including 5 μl of a solution containing l-lysine (1 mM) and l-methionine (1 mM), 5 μl of human urine, or a single colony of actively growing E. coli (of strains CFT073 or MG1655), added to the plate (7). In a soft-agar overlay containing 10⁴ quiescent CFT073 cells, we tested the known proliferants and in each case observed the development of a zone of actively growing colonies surrounding the site of challenge (Fig. 1B). Without challenge, cells failed to grow throughout the course of the experiment (24 h) and remained quiescent. We also tested additional clinical isolates and found that quorum-dependent growth, as measured by colony formation, on glucose minimal agar is common among the ST131 UPEC strains in this study (8/33) (Table 1); however, we discovered that the number of cells constituting a quorum varies among strains. For example, the ST131 strain JJ1886 was proliferative on glucose minimal agar at a plating density greater than 10³ CFU but quiescent at or below 10³ CFU (Table 1). Of the 35 strains tested in this study, 10 were reversibly quiescent on glucose minimal agar, and three of those strains were quiescent at or below a plating density of 10³ CFU but nonquiescent at 10⁵ CFU (Table 1). These results show that sensitivity for quorum detection varies by strain; i.e., it is strain specific.

FIG 1.

The quorum-regulated quiescent state of E. coli CFT073 cells on glucose M9 minimal agar is reversible. (A) CFT073 cells were seeded on 0.2% glucose M9 minimal agar plates and incubated for 24 h at 37°C at low density (10⁴ CFU), which promotes quiescence, and at high density (10⁸ CFU), which prevents quiescence and promotes proliferation, leading to a lawn of colony growth on the agar plate. (B) CFT073 cells (10⁴ CFU) were plated on M9 minimal agar with 0.2% glucose. A 5-μl solution of l-lysine (1 mM) and l-methionine (1 mM), sterile human urine, or an actively growing colony of CFT073 was transferred to the center of the plate, cells were incubated for 24 h at 37°C, and the zone of growth, measured by colony formation surrounding the challenge site, was monitored. (C) The zone of colony formation extended outward for 72 h after addition of the actively growing colony of CFT073 (proliferant) to the center of the plate containing quiescent cells. (D) The diameter of the zone containing colonies was measured and plotted against time. Data from three replicates are shown, with error shown as standard deviations. (E) Zonal expansion assay was repeated as for panel C, but the actively growing colony at the center of the plate (proliferant) was removed at 24 h. The diameter of the zone containing colonies continued to expand with time after removal of the initial proliferant, similar to the expansion measured in the presence of proliferant (D). Arrows indicate positions of proliferant addition, and brackets indicate the zone of colony growth. Bars, 1 cm. Images in panels A, B, C, and E are representative of at least three independent replicates.

TABLE 1.

Prevalence of quiescence among UPEC ST131 isolates

Strain	Sequence type	ST131 clade	Quiescence at 103 CFU and reversal by Lys and Meta	Reference or source
CFT073	ST73	NAd	+b	7
Nissle 1917	ST73	NA	+b	7
JJ1886	ST131	H30Rx/C2	+c	J. Johnson
JJ1887	ST131	H30Rx/C2	−	J. Johnson
JJ2547	ST131	H30Rx/C2	+c	J. Johnson
JJ2050	ST131		−	J. Johnson
JJ2528	ST131	H30R1/C1	−	J. Johnson
MVAST0036	ST131	H30R1/C1	−	J. Johnson
MVAST014	ST131	H30R1/C1	+c	J. Johnson
H17	ST131	H22/B0	−	J. Johnson
JJ1901	ST131		−	J. Johnson
JJ2087	ST131	H22/B	−	J. Johnson
JJ1999	ST131	H22/B	−	J. Johnson
MVAST167	ST131	H41/A	−	J. Johnson
MVAST020	ST131	H41/A	−	J. Johnson
JJ2134	ST131	H30Rx/C2	+b	J. Johnson
JJ2183	ST131	H30Rx/C2	−	J. Johnson
JJ2489	ST131	H30Rx/C2	−	J. Johnson
QUC02	ST131		−	J. Johnson
JJ2118	ST131	H30R1/C1	−	J. Johnson
JJ2193	ST131	H30R1/C1	−	J. Johnson
JJ2210	ST131	H30R1/C1	−	J. Johnson
JJ2578	ST131	H30R1/C1	−	J. Johnson
JJ2608	ST131	H30R1/C1	−	J. Johnson
MVAST038	ST131	H30R1/C1	−	J. Johnson
MVAST046	ST131	H30R1/C1	+	J. Johnson
MVAST077	ST131	H30R1/C1	−	J. Johnson
MVAST084	ST131	H30R1/C1	−	J. Johnson
MVAST131	ST131	H30R1/C1	−	J. Johnson
MVAST158	ST131	H30R1/C1	−	J. Johnson
MVAST179	ST131	H30R1/C1	+b	J. Johnson
JJ2546	ST131		−	E. Sokurenko
JJ2548	ST131		+b	E. Sokurenko
JJ2963	ST131		−	E. Sokurenko
JJ2974	ST131		+b	E. Sokurenko

^aStrains were seeded at 10³ CFU and 10⁵ CFU on 0.2% glucose M9 minimal agar plates and incubated at 37°C for 24 h. +, quiescence; −, nonquiescence.

^bStrain was quiescent when seeded at both 10³ CFU and 10⁵ CFU on 0.2% glucose M9 minimal agar and grown at 37°C for 24 h.

^cStrain was quiescent when seeded at 10³ CFU but nonquiescent when seeded at 10⁵ CFU on 0.2% glucose M9 minimal agar plates and grown at 37°C for 24 h.

^dNA, not applicable.

Next, we monitored the zone of colony formation of CFT073 cells as they exited quiescence over 72 h. First, we poured soft-agar plates containing quiescent CFT073 cells seeded at 10⁴ CFU per plate, placed an actively growing colony of CFT073 on the surface of each plate with a toothpick, and measured the zones of colony formation around the sites of inoculation at 24, 48, and 72 h. We observed that the zone of colony formation expands in diameter until the entire plate is covered by a lawn of actively growing cells (Fig. 1C to E). Two potential scenarios could explain this result. First, the proliferant(s) that stimulates colony formation of quiescent cells is continuously released from the toothpicked colony throughout the experiment and induces colony formation of distal quiescent cells as it diffuses across the plate. Alternatively, the proliferant(s) that stimulates colony formation from quiescent cells is released from the toothpicked colony, but upon release of the proliferant, neighboring and formerly quiescent cells begin secreting their own proliferant, thereby stimulating adjacent quiescent cells to grow. To distinguish between these possibilities, glucose minimal agar plates were seeded with 10⁴ CFT073 cells from a stationary-phase liquid glucose minimal medium culture, and a single CFT073 colony was toothpicked onto the center of each plate. The plates were then incubated for 24 h. During that time, a small zone of colony formation formed on each plate around the central toothpicked colony (Fig. 1C to E). After the zone had formed at 24 h, the toothpicked colony was removed to stop the colony from secreting additional proliferant(s). Even after removal of the toothpicked colony from the plate, the zone of colony formation continued to expand, unabated, for an additional 48 h and at a rate similar to that observed on the control plate, where the colony had been left intact (Fig. 1D and E). These results suggest that the proliferant(s) propagates across the plate from CFT073 cells as they exit quiescence. These results further show that quiescent CFT073 cells at the edge of the plate remain viable and can be stimulated to proliferate for up to 72 h.

Quiescent CFT073 cells are filamentous on glucose minimal agar plates.

During infection, UPEC cells from quiescent intracellular reservoirs (QIRs) display several different morphologies, including rods, spheres, and filaments (13–15). To determine if quiescent cells in our in vitro quiescence system also display morphological features that are distinct from those of actively growing cells, we compared the cell morphologies of CFT073 cells under several culture conditions, including LB agar and glucose minimal agar supplemented with l-lysine (1 mM) and l-methionine (1 mM) (Fig. 2A). After 24 h on LB agar or on glucose minimal agar supplemented with l-lysine and l-methionine, colonies were apparent, and individual cells collected and analyzed by microscopy appeared as short rods with mean lengths of 1.72 ± 0.59 μm (n = 200) and 2.33 ± 0.80 μm (n = 200), respectively (Fig. 2); no cells longer than 6.5 μm were detected. In contrast, after 24 h on nonsupplemented glucose minimal agar plates, no colonies were detected, and quiescent cells harvested from the plate had a broad length distribution, with a mean length of 7.17 ± 4.34 μm (n = 200), which is 300% greater than the mean length of cells grown on glucose minimal agar supplemented with lysine and methionine (Fig. 2). Forty-five percent of quiescent cells were longer than 6.5 μm, with 47.5 μm as the upper limit of the range, indicating a filamentous phenotype, and 16.5% of the filaments appeared to contain one or more incomplete septa (Fig. 2; see Fig. S1A in the supplemental material). To assess cells for length changes after extended incubation, we also measured cell length at 0 h and 48 h after entry into quiescence. At 0 h, cells harvested from a stationary-phase culture grown in liquid glucose minimal medium were short rods with a mean length of 2.14 ± 0.58 μm (n = 200). After 48 h on glucose minimal agar plates, quiescent cells were filamentous, with a mean length of 6.84 ± 3.76 μm (n = 72), which is similar to the mean length at 24 h (Fig. S1A and B). These results suggest that quiescent CFT073 cells elongate during the first 24 h on the plate but fail to complete division.

FIG 2.

Quiescent CFT073 cells are filamentous at 24 h. (A) CFT073 cells were grown on LB agar or 0.2% glucose minimal agar at 37°C with and without l-lysine (1 mM) and l-methionine (1 mM), harvested at 24 h, and visualized by DIC microscopy. Bars, 5 μm. (B) Violin plot representing cell length distribution at 24 h for 200 cells grown under conditions described for panel A.

Lysine import is necessary for lysine to prevent quiescence, but an additional proliferant is released from actively growing cells.

l-Lysine is a proliferant for quiescent CFT073 cells on glucose minimal agar in the presence of l-methionine (7). To determine whether l-lysine must be imported into CFT073 cells to prevent quiescence, we constructed a CFT073 lysP insertion-deletion mutant by λred recombination (Table 2). Of the three known lysine importers in E. coli, LysP is the major importer and a member of the amino acid, polyamine, and organocation (APC) family of transporters (16, 17). CFT073 ΔlysP cells grew well overnight in liquid glucose minimal medium, similar to the wild-type CFT073 strain (Fig. S2A), and exhibited quiescence on glucose minimal agar at low cell density (10⁴ cells per plate) (Fig. 3); however, it grew robustly when plated at high density (10⁸ cells per plate). Unlike wild-type CFT073, when 5 μl of a solution of l-lysine (1 mM) and l-methionine (1 mM) was added to quiescent CFT073 ΔlysP cells plated at 10⁴ CFU, no colony growth was observed after 24 h of incubation, suggesting that LysP-dependent transport is critical for lysine to stimulate proliferation. This also suggests that lysine biosynthesis is reduced in quiescent cells. CFT073 ΔlysP was complemented with a plasmid containing lysP (pLysP), which restored the ability of the strain to proliferate upon addition of l-lysine in the presence of l-methionine at low cell density (Fig. 3). Although quiescent CFT073 ΔlysP cells failed to respond to l-lysine, they proliferated after a colony of actively growing CFT073 cells was added to the plate (Fig. 3). Collectively, these results suggest that import of l-lysine is required for l-lysine to stimulate proliferation and override the absence of the quorum and that one or more l-lysine-independent proliferants that reverse quiescence are secreted from bacterial colonies.

TABLE 2.

E. coli strains and plasmids used for genetic analyses and constructions in this study

E. coli strain or plasmid	Relevant genotype description	Identifier	Reference, source, or construction
Strains
CFT073 Strr	Spontaneous streptomycin-resistant mutant of CFT073	CFT073	7
CFT073	Original clinical isolate	CFT073WT	45
MG1655	LAM-rph-1	MG1655	46
BL21(λDE3)	F− ompT gal dcm lon hsdSB(rB− mB−) λ(DE3[lacI lacUV5-T7 gene 1 ind-1 sam7 nin5])	BL21(λDE3)	EMD Millipore, USA
ED0052	CFT073 Strr ΔlysP::parE-kan	CFT073 ΔlysP	CFT073; λred
ED0070	CFT073 Strr ΔargP::cat	CFT073 ΔargP	CFT073; λred
ED0169	CFT073 ΔlysA::cat	CFT073 ΔlysA	CFT073WT; λred
JJ1886		JJ1886	47
ED0118	JJ1886 ΔmppA::cat	JJ1886 ΔmppA	JJ1886; λred
ED0131	JJ1886 ΔoppABCDF::cat	JJ1886 ΔoppABCDF	JJ1886; λred
Plasmids
pKD46	amp		43
pKD3	cat		43
pKD267	kan		B. Wannera; 48
pKD119	tet		43
pBAD24	amp (expression vector)		49
pLysP	amp Para::lysP	pLysP	This study
pArgP	amp Para::argP	pArgP	This study
pet-AmiD	kan (ΔN17)amiD-His6		This study

^aJ. Teramoto, K. A. Datsenko, and B. L. Wanner, unpublished construction.

FIG 3.

Quiescent CFT073 ΔlysP and ΔargP cells fail to proliferate with l-lysine. Strains CFT073, CFT073 ΔlysP, and CFT073 ΔargP were plated on M9 minimal agar with 0.2% glucose at 10⁴ CFU and then challenged with a 5-μl solution of l-lysine (1 mM) and l-methionine (1 mM), sterile human urine, or an actively growing colony of MG1655 transferred to the center of the plate. CFT073 ΔlysP cells containing pLysP or pBad24, where indicated, were supplemented with 50 μg ml⁻¹ ampicillin and 0.1% arabinose. CFT073 ΔargP cells containing pArgP or pBad24, where indicated, were supplemented with 50 μg ml⁻¹ ampicillin. Plates were incubated at 37°C for 24 h. Bar, 1 cm. Images are representative of three independent replicates.

ArgP is a transcriptional regulator of lysP (16, 18). In the absence of lysine, ArgP positively regulates transcription of lysP, along with several E. coli genes involved in lysine and PG biosynthesis, including dapB (dihydropicolinate reductase), dapD (tetrahydrodipicolinate succinylase), lysA (diaminopimelate decarboxylase), and lysC (aspartokinase III) (16, 18–20). To test whether ArgP plays a regulatory role in quorum-dependent colony formation or quiescence, we deleted argP. Like wild-type CFT073, CFT073 ΔargP cells grew well overnight in liquid glucose minimal medium (Fig. S2A) and were quiescent at low cell density (10⁴ CFU per plate) on glucose minimal agar (Fig. 3), indicating that ArgP is not essential for inducing quiescence at low density (Fig. 3). Similar to the result with CFT073 ΔlysP cells, quiescent CFT073 ΔargP cells failed to respond to the addition of l-lysine with l-methionine after 24 h of incubation, but the cells proliferated in response to a colony of CFT073 toothpicked to the center of a glucose minimal agar plate seeded with 10⁴ CFU of CFT073 ΔargP (Fig. 3). Complementation of CFT073 ΔargP with the wild-type CFT073 argP gene on a plasmid (pArgP) restored the ability of the strain to respond to l-lysine in the presence of l-methionine at low cell density (Fig. 3). Since LysP expression has been reported to be repressed 35-fold in E. coli argP mutants (20), this result both (i) supports the conclusion that external l-lysine must be transported into CFT073 via LysP in order to prevent quiescence and (ii) confirms that expression of LysP is regulated by ArgP. Together, these results also suggest that there is a molecule(s) other than l-lysine secreted by actively growing cells that stimulates proliferation of quiescent CFT073 cells. We previously reported that cell-free culture supernatants prepared from CFT073 grown in liquid glucose minimal medium contain l-lysine (7). Therefore, to confirm that it is not l-lysine secretion from the actively growing cells in the colony that reactivates quiescent cells and induces proliferation, we constructed a lysine-biosynthetic mutant of CFT073 by deletion of lysA (Table 2). This strain grows in glucose minimal medium supplemented with l-lysine but fails to proliferate on glucose minimal agar without lysine in response to an actively growing colony of CFT073 cells, which are capable of synthesizing, and therefore likely releasing, l-lysine (Fig. S3A). This is consistent with the suggestion that the molecule(s) secreted by cells that stimulates proliferation is not l-lysine.

To further clarify why the population threshold that defines a quorum varies between strains, we investigated the ST131 clinical isolate JJ1886, which exhibits quiescence on glucose minimal agar at or below a plating density of 10³ CFU but grows robustly on such plates at 10⁵ CFU and in liquid glucose minimal medium (Fig. 4A and Fig. S2B). We observed that 5 μl of a solution of l-lysine (1 mM) alone is sufficient to promote colony formation and prevent quiescence of JJ1886 plated at low density on glucose minimal agar (10³ CFU per plate), although colony formation is more efficient in the presence of l-methionine (1 mM), and l-methionine (1 mM) alone does not stimulate colony formation (Fig. 4B). Since quiescent JJ1886 cells on glucose minimal agar proliferate in the presence of l-lysine alone, we titrated the l-lysine concentration to determine the threshold for stimulation of colony formation and compared JJ1886 to CFT073. We determined that a low concentration of l-lysine (5 μl of 0.1 mM) is sufficient to promote proliferation of quiescent JJ1886 cells (10³ CFU per plate), whereas quiescent CFT073 cells (10³ CFU per plate) proliferated at a higher concentration of l-lysine (>0.5 mM) and only in the presence of l-methionine. These results suggest that although JJ1886 behaves phenotypically similarly to CFT073 with respect to quiescence, it is more sensitive to detecting or responding to an external proliferant(s) that prevents quiescence. We then tested if quiescent JJ1886 also responds to the other known CFT073 proliferants, including 5 μl of urine or a colony of actively growing E. coli cells, and observed that both agents were effective for stimulating proliferation of quiescent JJ1886 cells (Fig. 4B).

FIG 4.

ST131 strain JJ1886 displays quorum-dependent colony formation. (A) JJ1886 cells were plated on 0.2% glucose M9 minimal agar and LB agar at 10⁵ and 10³ CFU. JJ1886 displays quorum-dependent colony formation at 10³ CFU on glucose M9 minimal agar. (B) JJ1886 (10³ CFU) was plated on 0.2% glucose M9 minimal agar. A 5-μl solution of l-lysine (1 mM) and l-methionine (1 mM), l-lysine (1 mM), l-methionine (1 mM), sterile human urine, or an actively growing colony of CFT073 was transferred onto a JJ1886-seeded plate, and plates were incubated for 24 h at 37°C. Bars, 1 cm. Images are representative of at least three independent replicates.

Since l-lysine stimulates proliferation of both JJ1886 and CFT073, we tested 5 μl of d-lysine (1 mM) and 5 μl of the biosynthetic lysine precursor meso-diaminopimelic acid (DAP) (1 mM) for stimulating proliferation of quiescent cells. Both agents failed to stimulate colony formation from quiescent CFT073 and JJ1886 (Fig. S3B and C).

Peptidoglycan fragments prevent UPEC quiescence.

Next, we considered PG as a potential cue or signal that would communicate a quorum since PG fragments, including di-, tri-, and tetrapeptides, are normally secreted by actively growing bacterial cells, including E. coli (10, 12, 21, 22). We tested if peptidoglycan fragments from E. coli could prevent quiescence and stimulate proliferation of CFT073 and JJ1886. Peptidoglycan was digested with mutanolysin, which released a heterogeneous mixture of peptidoglycan fragments, including various muropeptides and stem peptides (Fig. 5A). Addition of the total mixture of digested peptidoglycan fragments from E. coli induced proliferation of quiescent JJ1886 cells, plated at 10³ CFU per plate (Fig. 5A), and quiescent CFT073 cells, plated at 10⁴ CFU per plate (Fig. S4A). Titrating the amount of total peptidoglycan added to the agar plate revealed a dose-dependent response, leading to differences in the zones of colony formation around the site of addition (Fig. S4B). Digested peptidoglycan from B. subtilis and Staphylococcus aureus also stimulated proliferation of quiescent JJ1886 and CFT073 cells (Fig. S4C and D). Furthermore, peptidoglycan fragments were also effective for stimulating proliferation of quiescent CFT073 ΔlysP cells, indicating that the peptidoglycan response is independent of lysine import (Fig. S5A). Interestingly, these results show that quiescent E. coli cells respond to PG fragments from various bacterial species, which can differ in the identity of the amino acid at the third position of the stem peptide, usually meso-DAP (mDAP) (i.e., E. coli) or l-lysine (i.e., S. aureus) (10, 23, 24).

FIG 5.

Peptidoglycan fragments promote proliferation of quiescent JJ1886 cells. LCMS total ion counts (TIC) are shown, with the corresponding test for quiescence reversal on the right. (A) Mutanolysin-digested crude E. coli PG, which was tested for stimulating proliferation of quiescent JJ1886 cells plated at 10³ CFU on 0.2% glucose M9 minimal agar. (B) Elution of a PG tripeptide (l-Ala–d-Glu–mDAP), [M+H]⁺ 391.275 (m/z), similar to an E. coli fragment. (C) Elution of a PG pentapeptide (l-Ala–d-Glu–l-Lys–d-Ala–d-Ala), [M+H]⁺ 489.380 (m/z), similar to an S. aureus PG pentapeptide. (D) Elution of a NAG-NAM disaccharide, [M+K]⁺ 535.128 (m/z). (E) Elution of a disaccharide tetrapeptide, [M+H]⁺ 940.396 (m/z), isolated from crude E. coli PG. (F) Products of the enzymatic reaction of E. coli-derived disaccharide tetrapeptide isolated from HPLC and then digested with AmiD to produce a tetrapeptide (l-Ala–d-Glu–mDAP–d-Ala), [M+H]⁺ 462.198 (m/z). Candidate proliferants were collected and tested in the quiescence assay as described in Materials and Methods, and plates were incubated at 37°C for up to 24 h. Arrows indicate positions of proliferant addition. Bars, 1 cm. Images of quiescence assay plates are representative of at least three independent replicates.

Peptidoglycan stem peptides act as proliferants for quiescent JJ1886 and CFT073 cells.

To identify which fragment(s) of digested PG stimulates proliferation of quiescent cells, we performed high-performance liquid chromatography-mass spectrometry (HPLC-MS) on E. coli PG fragment mixtures (Fig. 5A) and tested fractions across the elution for the ability to promote proliferation of quiescent JJ1886 cells. We found that early fractions that were collected with retention times up to 10 min were sufficient to stimulate proliferation of quiescent JJ1886 cells (Fig. S5B). Tandem mass spectrometry confirmed that these fractions contain low-molecular-weight peptidoglycan fragments, consistent with stem peptides, such as the tri- and tetrapeptides (Table S1). In contrast, later-eluting fractions containing muropeptides did not stimulate proliferation of quiescent JJ1886 cells (Fig. S5B).

To further explore the molecular determinants of PG peptides that stimulate proliferation of quiescent cells, we compared several high-purity PG peptide fragments of various amino acid compositions and lengths (Table S1). We observed that a tripeptide which could be generated from known amidase and carboxypeptidase activities in the E. coli periplasm (l-Ala–d-Glu–mDAP) failed to stimulate proliferation of quiescent JJ1886 cells (Fig. 5B). However, a peptidoglycan pentapeptide, l-Ala–d-Glu–l-Lys–d-Ala–d-Ala (A-E-K-A-A), which is similar to the stem pentapeptide used for S. aureus peptidoglycan synthesis, stimulated proliferation of quiescent JJ1886 (Fig. 5C) and CFT073 (Fig. S5A) cells. We also tested a smaller fragment of the pentapeptide, acetyl–l-Lys–d-Ala–d-Ala (K-A-A), but did not observe stimulatory activity (Fig. S6A). Fragments including the disaccharide moiety alone, containing NAG and NAM (NAG-NAM) (Fig. 5D), and higher-molecular-weight molecules with the attached stem peptide (i.e., disaccharide tetrapeptide) (Fig. 5E), in various states of cross-linking, eluted later, after 10 min (Fig. S7), and had no proliferant activity. NAG-NAG, NAG, and NAM were also tested, but none of these sugars stimulated proliferation of quiescent cells (Fig. S6B).

To determine if a peptidoglycan stem peptide longer than three amino acids and derived from E. coli stimulates proliferation of quiescent cells, we generated tetrapeptide in situ by digesting purified disaccharide tetrapeptide (isolated by HPLC) with AmiD, an amidase from E. coli (22). AmiD cleaves a stem peptide from attached sugars (Fig. S8A). After cleavage to release the tetrapeptide, the reaction products were highly active for stimulating proliferation of both quiescent JJ1886 cells (Fig. 5F and Fig. S8B) and quiescent CFT073 cells (Fig. S8C). Since NAG-NAM displays no activity alone, these results suggest that peptidoglycan peptides from E. coli (tetrapeptide) and S. aureus (pentapeptide) likely function as proliferants.

Finally, to further clarify if PG peptides exert their function without entering the cytoplasm or if they are imported into a quiescent cell and then enter the peptidoglycan recycling pathway (22), we constructed JJ1886 strains with deletions of the oligopeptide transporter operon, oppABCDF (25–29), and also the peptidoglycan tripeptide-specific delivery protein encoded by mppA (27, 30–32). Both deletion strains grew well in liquid glucose minimal medium and became quiescent when plated at 10³ CFU per plate, similar to parent strain JJ1886; furthermore, like JJ886, both strains were stimulated to proliferate by addition of stem peptides to the plate (Fig. S8B and D), indicating that it is not necessary for the stem peptides to enter the cytoplasm via the Opp transporter in order to act as a proliferant to overcome quiescence.

DISCUSSION

Here, we show that two virulent UPEC strains, CFT073 and JJ1886, enter a quiescent state on glucose minimal agar, where glucose is not limiting, and this state is prevented and reversed by proliferants. These proliferants, which prevent quiescence and promote colony formation, include peptidoglycan stem peptides consisting of four (e.g., E. coli) or five (e.g., S. aureus) amino acids, l-lysine and l-methionine, and human urine, which contains a high concentration of free amino acids (7). Quiescence is a slowly growing or nongrowing physiological state described for E. coli that is common among UPEC strains (i.e., 51/145 isolates tested) and is quorum regulated. The quiescent state is adopted by an entire population of cells under a given growth condition, in contrast to a persister state, which occurs in a small percentage of cells within a population (33–37). Adoption of a quiescent, nonproliferative state would require many physiological systems to slow (e.g., transcription and translation) or stop (e.g., DNA replication, division, and peptidoglycan synthesis). Accordingly, we observed that quiescent cells adopt a filamentous morphology, failing to successfully divide or proliferate despite extended incubation (see Fig. S1A in the supplemental material). Finally, different strains respond to known proliferants with various sensitivities, suggesting that either the sensitivity of detection or the regulatory control over physiological pathways varies among strains. For example, l-lysine is sufficient to induce proliferation of quiescent JJ1886 cells (Fig. 4B), whereas l-lysine must be supplemented with either l-methionine or l-tyrosine to induce proliferation of quiescent CFT073 cells (7). While lysine import is one way to overcome quiescence, stimulation by peptidoglycan stem peptides occurs independently of the lysine importer lysP, suggesting that there are multiple regulatory inputs that enable reinitiation of the cell cycle. It is also possible that detection of peptidoglycan fragments stimulates lysine biosynthesis or increases intracellular lysine concentration in a coordinated manner, which then induces proliferation.

Muropeptide fragments are known to resuscitate dormant Mycobacterium smegmatis (38) and to stimulate B. subtilis endospores to germinate (11). Here, we showed that in E. coli, PG stem tetra- and pentapeptide fragments not only cause cells to reenter the cell cycle but may also act as a signal or cue to communicate the presence of a quorum, independent of uptake by the Opp import system. Previously, it was shown that an oppA mutant is far less infective in a mouse UTI model than the wild type (39), and it was suggested that small oligopeptides, present in high concentration in urine (39), are imported and serve as major nutrients for CFT073 growth in urine (39). Here, neither the pentapeptide nor the tetrapeptide tested entered the quiescent cells via Opp, suggesting that they are imported by a second, presently unknown, oligopeptide permease, or act as signals in the periplasm and are recognized by a membrane receptor or signal transduction system that, when activated, promotes cell proliferation. The serine/threonine kinase PknB is a widely conserved PG binding protein, and its ortholog is important for germination of spores in B. subtilis (11); however, no similar proteins have been found thus far in E. coli.

The cause(s) of recurrent UTIs is complex (40); however, it appears that UPEC can bind to, enter, and replicate within superficial facet cells in the mouse and human bladder epithelium, resulting in intracellular biofilm-like communities (IBCs) (40, 41). IBCs escape from infected superficial facet cells within hours of development, and several cycles of IBC formation occur, but each successive round is associated with a reduced replication rate and smaller IBCs (15). The infected superficial facet cells exfoliate, exposing underlying transitional epithelial cells that become infected with IBC-derived UPEC progeny. Upon infection of transitional cells, small numbers of UPEC appear to enter a latent or quiescent intracellular state in endosomal vesicles (2 to 12 cells/vesicle), establishing QIRs (42) that are resistant to antibiotic treatment and are a possible cause of recurrent UPEC infection. If the physiology of UPEC quiescence in vitro is similar to that in vivo, understanding how PG stem peptides prevent in vitro quiescence at the molecular level may help inform how to prevent recurrent UPEC infections in humans.

MATERIALS AND METHODS

Bacterial strains and growth conditions.

E. coli strains and plasmids used in this study are listed in Table 2. Plasmids for complementation of both CFT073 ΔlysP and CFT073 ΔargP gene deletions were constructed using expression vector pBAD24. Both genes were amplified from CFT073 by whole-colony PCR (lysP, 5′ ATCAAGAATTCACATGGGTTCCGAAACTAAAACTACAGAAG 3′ and 5′ TAGATAAGCTTTTATTTCTTATCGTTCTGCGGGAACTTC 3′; argP, 5′ ATCAAGAATTCACATGCAACCCGACACAAAATTGTGTCATAG 3′ and 5′ TAGATAAGCTTTTAATCCTGACGAAGAACTTTGTGACCATAATCG 3′). The resulting PCR products were digested with EcoRI and HindIII. Digested PCR products were ligated into EcoRI/HindIII-digested pBad24. Both pLysP and pArgP plasmid constructs were verified by sequencing and then transformed into either CFT073 ΔlysP or CFT073 ΔargP deletion strains by electroporation. To test for complementation, transformants were grown overnight in 0.4% glucose M9 minimal medium with ampicillin (50 μg ml⁻¹) and were assayed for quiescence on 0.2% glucose minimal agar supplemented with arabinose (0.1%), where indicated.

The amiD gene sequence coding for residues 18 through 276 was amplified from E. coli MG1655, cloned into pET-24b using NdeI/HindIII restriction sites to incorporate a C-terminal hexahistidine tag (5′ ACTGACATATGGCAGGCGAAAAAGGCAT 3′ and 5′ GCAGTAAGCTTATCCTGCCCGTATTTC 3′), and expressed in BL21(λDE3). This resulted in removal of the N-terminal signal sequence and lipid modification site.

Clinical UPEC isolates were obtained from the collections of J. Johnson (29 strains) and E. Sokurenko (5 strains) (Table 1). Lennox broth (LB) supplemented with agar (LB agar) was used for routine cultivation. Liquid glucose (0.2% or 0.4%, where indicated) M9 minimal medium and 0.2% glucose M9 minimal agar plates were prepared as described previously (7). Gene deletions were constructed by λred recombination as described by Datsenko and Wanner and confirmed by sequencing (43). Culture density was monitored by growing cells in liquid 0.2% glucose M9 minimal medium and measuring optical density at 600 (OD₆₀₀) at the appropriate times.

Quiescence assay.

Quiescence assays were performed as described previously (7). Briefly, CFT073 and JJ1886 wild-type and mutant strains were grown on LB agar plates overnight at 37°C. An isolated colony was used to inoculate 5 ml of 0.4% glucose M9 minimal medium, incubated at 37°C with shaking overnight, and then diluted to between 10³ and 10⁸ CFU, where indicated, in 3 ml 0.2% glucose M9 minimal medium containing 0.75% Difco Noble agar held at 45°C (soft agar). Inoculated soft agar was immediately poured onto plates containing a 12-ml layer of 0.2% glucose M9 minimal medium with 1.5% Noble agar or directly into an 8.5-cm² polystyrene culture dish. Putative proliferants, including actively growing cells, amino acids, DAP (2,6-diaminopimelic acid), sterile human urine, and peptidoglycan fragments, were added, and plates were incubated at 37°C for 24 to 72 h, where indicated, and then imaged.

Microscopy.

Overnight cultures of CFT073 were grown to stationary phase in 0.4% glucose M9 minimal medium. Five microliters of a 3-log diluted culture was spotted onto LB agar or 0.2% glucose M9 minimal medium with 1.5% Noble agar and, where indicated, supplemented with l-lysine (1 mM) and l-methionine (1 mM) and was then allowed to dry for 1 h at room temperature. Plates were incubated at 37°C for 24 or 48 h. Cells were harvested onto a glass coverslip, resuspended in phosphate-buffered saline (PBS) with EDTA (1 mM) and applied to an agar gel pad containing 0.2% glucose M9 minimal medium with 0.5% Noble agar. Cells were visualized by differential interference contrast (DIC) microscopy using a Zeiss LSM 700 microscope, and images were captured on an AxioCam HRc high-resolution camera with ZEN 2012 software. Images were processed using Adobe Photoshop CC and analyzed using NIH ImageJ software.

Peptidoglycan fragment separation and analysis.

E. coli CFT073 was inoculated in 1 liter of M9 minimal medium containing 0.4% glucose and incubated for 24 h shaking (175 rpm). Methods used for PG isolation were adapted from reference 44. Briefly, cells were harvested at 5,000 × g for 10 min at room temperature, and pellets were resuspended in 3 ml of M9 minimal medium. Cell suspensions were transferred into 6 ml of 6% sodium dodecyl sulfate (SDS) and boiled for 3 h with stirring. After 3 h, heat was discontinued, and cell suspensions were stirred overnight at room temperature. Cell suspensions were centrifuged at 262,000 × g for 30 min, pellets were resuspended in liquid chromatography-mass spectrometry (LCMS) grade water, and further centrifugation and wash steps were carried out until no SDS remained. The final pellet was resuspended in 900 μl of 10 mM Tris-HCl with 0.06% (wt/vol) NaCl, 100 μg of pronase E was added to the cell suspension, and the mixture was incubated for 2 h at 60°C. After incubation, 200 μl of 6% SDS was added, and the mixture was boiled at 100°C for 30 min. Samples were centrifuged at 262,000 × g for 30 min and washed 4 times with LCMS grade water. The final pellet was resuspended in 200 μl of PBS and incubated at 37°C overnight with 200 U of mutanolysin (M9901; Sigma-Aldrich), which contains N-acetylmuramidase M1 (β-1,4-N,6-O-diacetylmuramidase) and contaminating amidase activity. Samples were centrifuged at 16,000 × g for 10 min to remove cellular debris, and PG supernatant was collected for further testing. Mutanolysin was removed by ultrafiltration using a sterile polyethersulfone filter with a 3-kDa molecular weight cutoff. Separation and initial chemical characterization were carried out by LCMS on a Thermo Fisher ISQ EC spectrometer coupled to a Shimadzu Prominence-i LC-2030 system. Chemical characterization of pure compounds was undertaken using electrospray ionization-mass spectrometry (ESIMS) using an AB Sciex TripleTOF 4600 spectrometer and LCMS/MS experiments.

Analytical high-performance liquid chromatography (HPLC) was used to purify the PG extract. HPLC experiments were performed on a Shimadzu Prominence-I LC-2030C system equipped with a PDA detector (model LC-2030/2040), pump (LC-2030), and autosampler (LC-2030). Preparative HPLC experiments were completed with a Waters XBridge analytical C₁₈ 5-μm, 4.6- by 250-mm column at 30°C using a solvent gradient of 10 to 100% B (5 mM ammonium acetate with 15% acetonitrile [pH 4.89]) run over 16 min against solvent A (5 mM ammonium acetate [pH 4.37]). Solvent B was held at 10% for 4 min prior to the gradient, and 100% B was held for a subsequent 4 min at the conclusion of the method. Individual peaks were collected according to absorbance values at 205 nm. LCMS/MS was performed on a AB Sciex TripleTOF 5600 spectrometer coupled to a Waters H class ultraperformance liquid chromatography (UPLC) system. The LC method mirrored that of initial screening with an adjusted flow rate of 0.5 ml min⁻¹ on an XBridge analytical C₁₈ 5-μm, 4.6- by 250-mm column. Acquisition (SWATH) was performed in positive-ionization mode. The method-specific parameters were as follows: gas 1 (GS1), 55 lb/in²; gas 2 (GS2), 60 lb/in²; and curtain gas (CUR), 25 lb/in². The source-specific parameters were as follows: temperature (TEM), 500°C; ion spray voltage floating (ISVF), 5,500 V; declustering potential (DP), 100; collision energy (CE), 10; and collision energy spread (CES), 15. An initial time-of-flight (TOF) scan was collected from m/z 100 to 1,500, and SWATH data were acquired (m/z 200 to 1,100) over 36 SWATH windows per cycle with a window size of m/z 25 and a m/z 1 overlap between windows.

Peptidoglycan from B. subtilis (69554; Sigma-Aldrich) or S. aureus (77140; Sigma-Aldrich) (10 mg ml⁻¹) was incubated with 200 U of mutanolysin at 37°C overnight. Peptidoglycan stem peptide fragments, including the A-E-K-A-A pentapeptide (l-Ala–d-Glu–l-Lys–d-Ala–d-Ala) (A0910; Sigma-Aldrich), K-A-A tripeptide (acetyl–l-Lys–d-Ala–d-Ala) (A6950; Sigma-Aldrich), and A-E-DAP tripeptide (l-Ala–d-Glu–mDAP) (AS-60774; AnaSpec), and NAG-NAM (A178230; Toronto Research Chemicals) were each tested at 5 μl (2 mM) against strains JJ1886, JJ1886 ΔoppABCDF, and JJ1886 ΔmppA (10³ CFU) after growth on M9 minimal agar with 0.2% glucose for 24 h at 37°C.

Amidase cleavage of disaccharide tetrapeptide.

E. coli BL21(λDE3) cells containing expression vector pET-AmiD, modified to remove 17 amino acids from the N terminus of AmiD, were grown in LB supplemented with 50 μg ml⁻¹ kanamycin at 37°C to an OD₆₀₀ of 0.8. Expression was induced with isopropyl-β-d-thiogalactoside (1 mM) for 3 h at 30°C. Cells were harvested by centrifugation at 6,000 × g for 30 min at 4°C and resuspended in 25 ml of 25 mM Tris-HCl (pH 7.8) buffer containing 150 mM NaCl, 1 mM MgCl₂, and 10% glycerol. Cells were lysed with a French press. Soluble cell extracts were collected by centrifugation at 30,000 × g for 30 min at 4°C, then applied to Talon Superflow resin (GE Healthcare), washed with 10 column volumes of lysis buffer, and eluted with an imidazole gradient. Buffer was exchanged by a PD-10 desalting column to remove imidazole. Isolated E. coli PG disaccharide tetrapeptide was digested with 315 ng of AmiD in 8.5 mM Tris-HCl (pH 7.8), 50 mM NaCl, 0.3 mM MgCl₂, and 3% glycerol for 26 h at 37°C. Cleavage was confirmed by LCMS analysis as described above.