LiaX is a surrogate marker for cell envelope stress and daptomycin non-susceptibility in Enterococcus faecium

ABSTRACT

Daptomycin (DAP) is often used as a first-line therapy to treat vancomycin-resistant Enterococcus faecium infections, but emergence of DAP non-susceptibility threatens the effectiveness of this antibiotic. Moreover, current methods to determine DAP minimum inhibitory concentrations (MICs) have poor reproducibility and accuracy. In enterococci, DAP resistance is mediated by the LiaFSR cell membrane stress response system, and deletion of liaR encoding the response regulator results in hypersusceptibility to DAP and antimicrobial peptides. The main genes regulated by LiaR are a cluster of three genes, designated liaXYZ. In Enterococcus faecalis, LiaX is surface-exposed with a C-terminus that functions as a negative regulator of cell membrane remodeling and an N-terminal domain that is released to the extracellular medium where it binds DAP. Thus, in E. faecalis, LiaX functions as a sentinel molecule recognizing DAP and controlling the cell membrane response, but less is known about LiaX in E. faecium. Here, we found that liaX is essential in E. faecium with an activated LiaFSR system. Unlike E. faecalis, E. faecium LiaX is not detected in the extracellular milieu and does not appear to alter phospholipid architecture. We further postulated that LiaX could be used as a surrogate marker for cell envelope activation and non-susceptibility to DAP. For this purpose, we developed and optimized a LiaX enzyme-linked immunosorbent assay (ELISA). We then assessed 86 clinical E. faecium bloodstream isolates for DAP MICs and used whole genome sequencing to assess for substitutions in LiaX. All DAP-resistant clinical strains of E. faecium exhibited elevated LiaX levels. Strikingly, 73% of DAP-susceptible isolates by standard MIC determination also had elevated LiaX ELISAs compared to a well-characterized DAP-susceptible strain. Phylogenetic analyses of predicted amino acid substitutions showed 12 different variants of LiaX without a specific association with DAP MIC or LiaX ELISA values. Our findings also suggest that many E. faecium isolates that test DAP susceptible by standard MIC determination are likely to have an activated cell stress response that may predispose to DAP failure. As LiaX appears to be essential for the cell envelope response to DAP, its detection could prove useful to improve the accuracy of susceptibility testing by anticipating therapeutic failure.

INTRODUCTION

Vancomycin-resistant enterococci (VRE) are organisms that cause serious healthcare-associated infections (1). Enterococcus faecium accounts for greater than 75% of VRE (2), and a concerning number develop resistance to additional antibiotics, exhibiting the multidrug-resistant phenotype. Daptomycin (DAP), a cyclic bactericidal lipopeptide, has become a first-line choice to treat severe VRE infections. The mechanism of DAP action involves insertion in the membrane of the calcium-bound antibiotic molecule, where it interacts with phosphatidylglycerol and lipid II intermediates to disrupt cell wall biosynthesis and cell membrane homeostasis (3). However, after introduction of DAP in clinical practice, DAP resistance (DAP-R) has been increasingly documented. Indeed, some data suggest that DAP-R occurs in up to 20%–40% of VRE bloodstream infection isolates in some institutions (4, 5) and can emerge during therapy (6, 7), jeopardizing the clinical utility of this antibiotic against recalcitrant VRE infections. Moreover, DAP-R has also been documented in isolates recovered from patients who have never been exposed to the antibiotic (8).

The LiaFSR system is a three-component cell envelope stress regulatory system that has been implicated in DAP resistance both in Enterococcus faecalis and E. faecium. Deletion of the genes encoding the response regulator LiaR in both species has been shown to cause hypersusceptibility to DAP (9, 10). Characterization of the LiaR regulon in both E. faecalis and E. faecium indicated that the main target of the response regulator is a cluster of genes encoding three proteins designated LiaX, LiaY, and LiaZ (11). LiaX was initially characterized as a protein that binds penicillin-binding protein (PBP) 5 in E. faecium and is likely involved in β-lactam resistance (12). Subsequently, in E. faecalis, it was found that LiaX is a surface-exposed protein harboring two distinct domains with specific and independent functions. Indeed, the C-terminal domain of LiaX functions as a regulator of LiaFSR by negatively controlling phospholipid remodeling that occurs during the “attack” by DAP and antimicrobial peptides, a phenotype that is an indicator of cell membrane adaptation in E. faecalis. On the other hand, the N-terminus appears to be released from the surface during cell envelope stress and functions as a sentinel signal transduction molecule by binding to DAP and antimicrobial peptides, a complex that is likely to be recognized on the cell surface to maintain the cell membrane stress response, critical for cell survival (13). In E. faecium, however, the genomic context of the liaXYZ gene cluster differs compared to E. faecalis, and, most importantly, the major membrane changes in anionic phospholipid microdomains associated with development of DAP resistance by E. faecalis do not seem to occur in E. faecium.

Establishing DAP susceptibility in clinical strains of E. faecium has been challenging, mainly because choosing a breakpoint is complicated by the wild-type distribution of the isolates with many exhibiting higher minimum inhibitory concentrations (MICs) than E. faecalis and falling beyond the breakpoint initially proposed. This situation has prompted the Clinical and Laboratory Standards Institute (CLSI) to change the DAP breakpoints (14) and establish different values for E. faecalis and E. faecium. Moreover, DAP MIC determination is influenced by many variables including the calcium concentration and the specific lot of Mueller-Hinton used by clinical laboratories. Furthermore, reproducibility of broth microdilution and available Etests is questionable, with data suggesting poor interlaboratory agreement when performing MIC of the same isolates (15). In this scenario, the use of DAP for severe VRE infections becomes difficult, particularly when susceptibility testing is not accurate and its reliability is dubious. This situation puts patients at risk for poor outcomes, particularly those who are often immunocompromised. Of note, while CLSI refers to E. faecium strains with DAP MICs ≤4 µg/mL as susceptible dose dependent (SDD), we will use the term DAP-susceptible (DAP-S) for ease of presentation.

In this work, we hypothesized that LiaX levels could be used to better assess DAP susceptibility in clinical E. faecium strains compared to standard MIC determination. Since standard susceptibility testing for DAP has major limitations in reproducibility and consistency (15), detection of a surrogate marker of cell membrane adaptation that accurately reflects susceptibility to DAP in E. faecium could serve as a more reliable tool to detect DAP resistance or tolerance and may potentially be clinically useful. During this study, we also provide further insights into the role of liaX in the cell envelope stress response of E. faecium.

RESULTS

LiaX in E. faecium is not detected in the extracellular milieu

In E. faecalis, LiaX is readily identified in the extracellular milieu, where it is thought to bind to incoming calcium-bound DAP molecules and antimicrobial peptides, as a mechanism for sensing incoming “threats.” Thus, we sought to investigate if LiaX could also be readily detected in the extracellular medium in E. faecium. To characterize the localization and dynamics of LiaX in E. faecium, we generated anti-LiaX antibodies for use in immunoblotting and enzyme-linked immunosorbent assays (ELISAs). We overexpressed full-length LiaX in E. coli from the well-characterized DAP-R E. faecium reference strain R494 (16). Subsequent purification resulted in an aliquot of LiaX protein assessed to have a purity of >95% (Fig. S1). Using the purified LiaX protein, we raised anti-LiaX antibodies in goats, which subsequently underwent high-affinity purification against LiaX.

To assess the analytical specificity of the purified polyclonal antibodies (Abs), immunoblotting was performed on E. faecium strains which harbored deletions of liaR and/or liaX. We generated mutants of liaR (the gene encoding the response regulator of the LiaFSR system) in a commensal strain of E. faecium (TX1330RF) (17). Deletion of liaR (TX1330RFΔliaR) is expected to abolish transcription of the liaXYZ gene cluster (9–11, 13). Furthermore, we attempted to delete liaX in TX1330RF but were only able to delete this gene in the background of TX1330RFΔliaR, generating a double mutant (TX1330RFΔliaRΔliaX) (see below). Using these strains, we were able to detect bands at the expected molecular weight of E. faecium LiaX (58.6 kDa) with purified LiaX and cell lysates of TX1330RF and TX1330RFΔliaR, but not with TX1330RFΔliaRΔliaX, indicating robust specificity of the primary polyclonal Ab for LiaX (Fig. 1A). These results also suggested that a basal amount of LiaX is produced even in absence of the LiaR response regulator.

Fig 1.

Immunoblots of E. faecium LiaX protein, mutant E. faecium strains, and clinically derived E. faecium isolates. (A) We were able to detect LiaX in TX1330ΔliaR even in the absence of LiaR. The LiaX signal is abolished in TX1330ΔliaRΔliaX. Immunoblots of cell lysates of TX1330 and derivative strains show bands present at the expected molecular weight of E. faecium LiaX (58.6 kDa). (B) LiaX is detected in higher amounts in DAP-R compared to DAP-S clinically derived E. faecium isolates and is not detected in extracellular media. Immunoblots of cell lysates and protein-precipitated extracellular media (supernatants) demonstrate increased LiaX production in DAP-R R497 and R494 compared to DAP-S S447 cell lysates. Supernatants of the same strains showed no detectable LiaX. The numbers indicate the molecular weights of bands in kilodaltons (kDa). RNA polymerase β-subunit was used as loading control.

Next, using the specific anti-LiaX antibodies, we sought to investigate the localization of LiaX in E. faecium isolates. Of note, in clinical and laboratory strains of E. faecalis, LiaX can be detected on the cell surface and in the extracellular media (N-terminal domain). Immunoblotting was performed on well-characterized clinical strains of DAP-S (S447) (6, 16, 18) and DAP-R E. faecium (R497 and R494) (16) using purified polyclonal goat antibodies. Strains were grown to mid-exponential phase; then, cells were separated from media for further analysis. Cells were lysed by glass beads, while the media underwent filter sterilization, and proteins were precipitated by the addition of trichloroacetic acid (TCA). Immunoblotting was performed on all samples after normalization for protein concentration. Quantification of the LiaX protein in cell lysates from DAP-R R497 and R494 was performed using Western blot analysis. Most importantly, we were unable to detect LiaX in the supernatants of any strain, suggesting that, unlike E. faecalis, LiaX in E. faecium is not released to the extracellular milieu (Fig. 1B). Furthermore, we confirmed that media supernatant spiked with purified LiaX protein (twofold dilution, range 2.5–0.019 ng) was detectable by Western blotting, with a limit of detection of LiaX at approximately 0.3 ng (Fig. S2).

Performance and optimization of the LiaX ELISA in clinical strains of E. faecium

We then optimized sensitivity of the purified polyclonal Ab for use in an ELISA by conducting a dilution titer. The purified Ab resulted in higher A₄₀₅ at every Ab dilution compared to unpurified Ab. Simultaneously, there was absence of off-target or “background” signal (A₄₀₅ = 0) from the goat pre-immune sera (Fig. 2), indicating the absence of immunoreactivity to LiaX prior to immunization with LiaX. The ELISA titer was performed with a dilution of commercial enzyme-conjugated secondary Ab recommended by the manufacturer, which had previously been optimized in enterococcal bacterial studies (1:2,000 or 0.3 µg/mL) (19). The titer was also performed with a constant antigen concentration of 10 ng of LiaX protein, which was representative of the A₄₀₅ that was observed from ELISA performed on E. faecium strains in preliminary studies. The purified primary Ab dilution to produce the highest and most discriminatory signal for ELISA was determined to be 1:400 (Ab = 4.185 µg/mL).

Fig 2.

Titer of multiple primary detection agents against LiaX. ELISA with 10 ng E. faecium LiaX antigen coated per well. Results are shown in triplicate and plotted with mean and SEM. Not all error bars are visible due to small size.

We subsequently performed a checkerboard titration of primary Ab and enzyme-conjugated secondary Ab using four different E. faecium strains (DAP-S strains S447 and 503 and DAP-R strains R446 and C1547). These strains were known to yield A₄₀₅ values for the LiaX ELISA from the minimum to maximum spectrophotometer detection limits of visible assay absorbance from preliminary studies (Fig. S3). The dilutions that were found to yield the A₄₀₅ values within the optimal accuracy range of our spectrophotometer with the greatest resolution were 1:400 for primary Ab and 1:2000 for secondary Ab. Using these concentrations, there was no off-target detection in the blank controls, indicating a highly specific assay.

After optimization by systematic checkerboard titration, the ELISA was performed on a collection of previously well-characterized, whole genome-sequenced, clinically derived E. faecium (20) with a wide range of DAP MICs (Table S1). The LiaX ELISA was able to differentiate between DAP-R and DAP-S strains using a preliminary cut-off at the A₄₀₅ level based on DAP-S strain S447 (Fig. 3). Assays of reference strains were performed in four different iterations (four different microtiter plates) over a 6-month period, with intra-assay coefficient of variation (CoV) ranging from 0.2% to 9.0% and overall inter-assay CoV ranging from 7.4% to 23.8% (Table S2). These observations fell within our pre-determined range of acceptable variation (highest inter-assay CoV of approximately 20%–25%, and not exceeding 30%) for the LiaX ELISA (21).

Fig 3.

Whole cell indirect LiaX ELISA of E. faecium reference and deletion strains using goat purified polyclonal Ab. All DAP-R strains (black boxes) have significantly higher LiaX levels than DAP-S strains (gray boxes). E. faecium TX1330ΔliaRΔliaX, with a double deletion of liaX and liaR, does not show any difference in LiaX detection compared to TX1330ΔliaR, which has only a deletion of liaR. DAP MICs are shown at the top of the chart for each strain,. *P < 0.05, **P < 0.001 compared to S447.

Once the validation step was carried out in well-characterized E. faecium strains, we performed the ELISA LiaX test on a collection of 86 clinical strains of E. faecium. The E. faecium isolates had been collected from a multicenter global cohort of patients with enterococcal bacteremia—the Vancomycin-Resistant Enterococcal Bloodstream Infection Outcomes Study (VENOUS) (22). We chose the isolates based on the availability of whole genome information. We performed standardized gradient strip and broth microdilution (BMD) assays to determine DAP MICs on all 86 E. faecium clinical isolates. DAP MICs were overall higher when performed by gradient strip compared to BMD (by a median twofold higher, interquartile range [IQR] 1.5–3). Of note, there was no disagreement in interpretive categories (i.e., SDD versus R) between gradient strip and the BMD method using the current CLSI interpretations (data not shown). DAP MICs determined by BMD were plotted against the LiaX ELISA A₄₀₅ on a receiver operating characteristic (ROC) curve, and the cut-off was determined by maximizing the Youden index (23). Of note, a high proportion of E. faecium isolates (n = 64; 74%) in the VENOUS cohort exhibited an elevated LiaX ELISA A₄₀₅ using the established cut-off (Fig. 4). We only had three E. faecium isolates with confirmed MICs within the resistant range (≥8 µg/mL), and all exhibited a LiaX ELISA well above the cut-off value. Most importantly, we had a total of 61 out of 83 (73.5%) E. faecium isolates with DAP-S MICs (DAP MIC ≤4 µg/mL) that exhibited LiaX ELISA A₄₀₅ values above the cut-off. When categorizing the results of the ELISA by MIC determined by BMD, more isolates with MIC <2 µg/mL had the LiaX ELISA A₄₀₅ above the cut-off compared to isolates with MICs 2–4 µg/mL (Fig. 4). This situation was reversed when using gradient strips with more isolates with MICs 2–4 µg/mL exhibited a high LiaX ELISA A₄₀₅ compared to those with MICs <2 µg/mL.

Fig 4.

Detection of LiaX by ELISA in daptomycin-resistant and daptomycin-susceptible E. faecium. Minimum inhibitory concentrations (MICs) by both (A) broth microdilution and (B) gradient strip tests.

Next, we evaluated the LiaX amino acid sequences of the 86 clinical isolates and eight clinical strains used as references to assess sequence variations that may affect the performance of the LiaX ELISA. The LiaX amino acid sequences were compared to the reference sequence of E. faecium TX16 (“Efm DO”). The majority (59%) of the E. faecium clinical isolates had a LiaX amino acid sequence with 100% identity to that of E. faecium DO. There were 11 variants of LiaX, with changes encompassing 1 to 11 amino acid substitutions (Table S3) compared to the reference sequence. There were no non-sense or frameshift alterations detected. Phylogenetic analysis of the LiaX variants showed a main cluster of LiaX variant 1 through 7 and a separate cluster of LiaX variant 8 through 12 (Fig. 5). We did not find any statistically significant association between any particular LiaX variant and discrepant DAP MIC/LiaX ELISA (i.e., DAP-S MIC with elevated LiaX ELISA) results. These findings suggest that the LiaX amino acid sequences identified in this study do not affect binding of the primary anti-LiaX antibody in the LiaX ELISA and, therefore, performance of the assay. Additionally, no association was found between LiaX variants and DAP MICs or LiaX ELISA A₄₀₅ overall.

Fig 5.

Phylogenetic analysis of LiaX sequences from E. faecium using a maximum likelihood tree.

LiaX is essential in E. faecium in strains with an intact LiaFSR system

In E. faecium, LiaX is a 522-amino acid protein similar in primary sequence to LiaX of E. faecalis, harboring an N-terminal domain predicted to be mainly α-helices and a C-terminus composed of β-pleated sheets (Fig. S4). To gain insights into the LiaR regulon in E. faecium, we had previously (11) performed a transcriptional mapping of genes regulated by LiaR in the E. faecium DAP-R clinical strain R497 and its liaR deletion mutant derivative (R497ΔliaR). As mentioned above, we expected that deletion of liaR would abolish liaX expression. Indeed, the genes of the liaXYZ operon were downregulated in the R497ΔliaR mutant (11). We confirmed these results by quantitative real-time polymerase chain reaction (qRT-PCR), which demonstrated downregulated expression of liaFSR/liaXYZ in the liaR deletion mutant (Fig. S5), supporting the notion that liaXYZ were under the control of LiaR in E. faecium clinical strains.

Subsequently, we attempted to delete liaX in the clinical strain DAP-R R497. Several attempts for mutagenesis were unsuccessful. We then used a different strain of E. faecium (TX1330RF), a clade B isolate that is a gut colonizer from the gastrointestinal tract of a human (17), which we have previously used for targeted mutagenesis (24, 25). Of note, TX1330RF is a DAP-susceptible, rifampin- and fusidic acid-resistant derivative of TX1330. Using a markerless system (26), we attempted deletion of liaX, and although we obtained first-event recombination integrants, we were unable to obtain a null liaX mutant. After multiple attempts using different strains employing the PheS* counterselection system, we were unable to recover any liaX mutants in TX1330RF.

As manipulation of liaX proved challenging in both DAP-S and DAP-R E. faecium strains, we hypothesized that LiaX was an important protein and that a basal level of this protein was necessary to maintain cell envelope integrity and response to cell surface stressors through the action of the LiaFSR system. Thus, we postulated that deletion of liaX might be possible in strains which have a non-functional LiaFSR system. This notion was supported by the fact that we had previously deleted liaR in several E. faecium strains (9, 10). Thus, we generated a liaR deletion mutant in TX1330RF prior to attempting liaX mutagenesis. Using this strategy, we were able to create a liaX deletion mutant in the TX1330RFΔliaR background (TX1330RFΔliaRΔliaX). Since TX1330RFΔliaR is DAP hypersusceptible (MIC = 0.125 µg/mL) compared to its parent TX1330RF (Table S1), a pattern previously observed in other E. faecium strains upon deletion of liaR (9), deletion of liaX did not have any effect on the DAP MIC (Table S1), supporting the assumption that LiaX is only essential within the context of a functional LiaFSR system. Of note, LiaX was originally discovered in E. faecium bound to PBP5 and was found to be associated with cephalosporin resistance in certain E. faecium without functional class A PBPs (12). Ceftriaxone (CRO) MICs of wild-type TX1330RF and deletion mutants were all 512 µg/mL.

Next, we attempted to complement liaR into TX1330RFΔliaRΔliaX to obtain a strain which functionally had only a liaX deletion. Efforts to complement liaR in TX1330RFΔliaRΔliaX in its native chromosomal location were unsuccessful, indicating that an active liaR affects cell viability in the absence of LiaX. Ultimately, we attempted a conditional complementation of liaR using the nisin-controlled vector pMSP3535 and successfully generated a derivative strain [TX1330RFΔliaRΔliaX (pMSP3535::liaR)] that was viable. However, the strain showed a major delay in growth of ~3 hours compared to its parental TX1330RF (Fig. S6), supporting the notion that the absence of liaX, in the background of a functional LiaFSR system, causes major biological fitness costs. These results also suggest that LiaX in E. faecium functions as a major modulator of the LiaFSR response, similar to E. faecalis, where the C-terminal domain of LiaX behaves as an inhibitor of the LiaFSR system (13).

Cell membrane phospholipid architecture and LiaX ELISA of E. faecium liaX and liaR mutants

In E. faecalis, activation of the cell envelope response to DAP and antimicrobial peptides via the LiaFSR system is accompanied by major changes in anionic phospholipid distribution (13, 27). Specifically, anionic phospholipid microdomains migrate away from the septum in an attempt to divert positively charged DAP from critical septal areas and prevent major damage during cell division and viability. Using 10-N-nonyl acridine orange (NAO) staining, we have previously (9) evaluated the architecture of anionic phospholipid microdomains in clinical strain pairs of DAP-S and DAP-R E. faecium and found that development of DAP-R is not associated with changes in phospholipid redistribution. As expected, deletion of liaR or both liaX and liaR in TX1330 also did not produce any changes in phospholipid microdomain distribution compared to parental wild-type strain (Fig. S7).

Finally, using our LiaX ELISA assay, we evaluated LiaX presence in the above mutants. Deletion of liaR in E. faecium strain TX1330RF (TX1330RFΔliaR) led to a decrease in the A₄₀₅ of the LiaX ELISA to a basal level compared to the parental (Fig. 3). Deletion of liaX in TX1330RFΔliaR did not lead to any further decrease in the A₄₀₅ of the LiaX ELISA. The ELISA signal in these two mutants is likely due to the low background signal of immunoglobulins binding to the plastic surface (28).

DISCUSSION

DAP is considered a bactericidal first-line antibiotic against vancomycin-resistant E. faecium despite lacking FDA indication for this purpose, particularly in highly immunosuppressed patients where bactericidal therapy is often preferred. A major issue of using DAP against vancomycin-resistant E. faecium is the emergence of resistance during therapy. Furthermore, there is evidence that DAP resistance can emerge even in the absence of DAP exposure (8), suggesting that the mechanism of protection against the antibiotic likely involves a major strategy that responds to DAP and other cell envelope stressors, including antimicrobial peptides. Indeed, the LiaFSR system appears to be a “broad” cell envelope stress system that can be readily activated in the presence of antibiotics and other cell envelope stressors. The LiaFSR system of E. faecium has similar target genes as those of E. faecalis, namely, liaXYZ. However, there are major differences in the mechanistic strategy underlying DAP-R between E. faecalis and E. faecium. Indeed, the main global strategy to counteract the effect of DAP in E. faecalis appears to be to divert the antibiotic from the septum to other areas of the cell membrane. This phenomenon involves major changes in the cell envelope, including redistribution of anionic phospholipid microdomains (27). In contrast, E. faecium seems to “repel” DAP from the cell surface in a mechanism that likely involves changes in surface charge without major alterations in phospholipid architecture (20). Regardless of the mechanism, the LiaFSR system-mediated response seems to be a key factor to successfully survive in the presence of DAP in both enterococcal species.

LiaX emerges as a critical effector of the LiaFSR system playing a key role at the cell membrane and envelope level. This protein was initially described (12) in E. faecium bound to PBP5 (designated Pbp5-associated protein, P5AP). Indeed, P5AP (LiaX) was found to be a mediator of cephalosporin resistance in E. faecium whose typical class A PBPs are not functional. Moreover, the expression of p5ap (liaX) was noted to be influenced by the activity of the serine-threonine phosphatase/kinase system (12). Consistent with earlier studies and our transcriptional and ELISA experiments, the main regulator of liaX is LiaR. However, other factors could influence its expression since we were able to detect LiaX even in absence of a functional LiaR regulator, confirming those initial observations that this key protein can be regulated by different molecular systems. Although the function of LiaX in E. faecium is not fully understood, its association with PBPs and our mutagenesis experiments suggest that liaX is essential during cell envelope stress and indicates an even broader role for LiaX in adapting the cell membrane to the DAP “attack.”

We have previously shown (13) that, in E. faecalis, DAP-R clinical isolates exhibited elevated LiaX levels, as detected by ELISA. Moreover, determination of DAP MICs using standard methodologies seem to be inaccurate and not reproducible between clinical microbiology laboratories, even in settings with known expertise in performing MIC determinations (15). Furthermore, there are well-documented cases of failure of DAP monotherapy (7, 29) in patients whose isolates were reported “susceptible.” The challenges in DAP MIC testing are derived from the fact that variations are observed in the media lots and calcium concentrations, making the standardization of the test very difficult (30). Therefore, novel tools that could predict susceptibility to DAP more accurately are clearly needed.

Based on our initial observations and under the hypothesis that LiaX is a key protein in the presence of cell envelope targeting antimicrobials in enterococci, we evaluated the possibility that LiaX may be used as a more accurate surrogate marker of cell envelope stress and, therefore, DAP susceptibility in E. faecium. Thus, we first developed and optimized an ELISA method for LiaX, assessing its performance on well-characterized strains of E. faecium. Consistent with our hypothesis, DAP-R strains indeed had significant higher levels of LiaX compared to DAP-S clinical strains (similar to E. faecalis). Furthermore, using an established cut-off that resulted from initial experiments, we could identify all DAP-R strains with MIC ≥8 µg/mL. Subsequently, to validate our findings, we applied our LiaX ELISA method to 86 E. faecium isolated from bacteremia recovered from patients in an ongoing large cohort study of enterococcal bacteremia (VENOUS) (22). The isolates were selected based on a convenience sample from sites in Houston and Detroit participating in VENOUS and mostly recovered from patients with hematological malignancies and solid organ transplant recipients.

As expected, all DAP-R isolates within our validation cohort with MICs ≥8 µg/mL exhibited elevated LiaX levels supporting the correlation between an upregulated cell membrane stress and DAP non-susceptibility. Most importantly, 73% of E. faecium reported as “susceptible dose-dependent” exhibited LiaX levels above the established cut-off. This finding suggests that a large majority of invasive E. faecium (all VENOUS isolates were recovered from the blood of included patients) have already triggered membrane adaptation changes that are likely to affect DAP susceptibility. The “stressors” are likely to include antimicrobial peptides produced by the innate immune system during the process of gut colonization and infection within the human host. These findings also provide some explanation of why some E. faecium strains that had never been exposed to DAP are non-susceptible at the first encounter with the antibiotic.

The discrepancies between activation of cell membrane stress response and DAP susceptibility have been previously described by our group in E. faecalis (31). Indeed, using a clinical strain pair of DAP-S and DAP-R E. faecalis recovered from a patient with a fatal bacteremia (16), we showed that introduction of a single amino acid substitution in LiaF (a member of LiaFSR system) resulted in activation of cell membrane remodeling and DAP tolerance in time-kill assays despite minimal changes in the DAP MIC (31). Our results in this work add more evidence to the fact that DAP MIC determination has major limitations to predict actual susceptibility to this antibiotic and support our premise that more accurate tools are critically required.

Using whole genome sequencing, we explored the possibility that variations in LiaX detection with our ELISA test were due to sequence polymorphisms. We did not find any correlation between specific sequence variants and the results of the ELISA or MICs, suggesting that the amino acid substitutions seen in LiaX in our heterogenous population of E. faecium acquired from patients with bacteremia do not affect antibody binding to LiaX, although we cannot rule out differences in liaX expression between the isolates that may affect our ability to detect LiaX. These initial results are encouraging and suggest that studies with more specific antibodies (e.g., anti-LiaX monoclonal antibody) could further improve the sensitivity and specificity of the LiaX ELISA, eliminating concerns of potential alterations of LiaX structure due to substitutions.

In summary, LiaX in E. faecium seems to be important for cell viability upon activation of cell envelope stress mediated by members of the LiaFSR system. Although the exact function of LiaX in E. faecium remains to be fully characterized, its involvement in cell membrane and PBP homeostasis suggests that this protein plays a critical role in defending E. faecium from cell envelope-acting antimicrobials. Using an ELISA method, we could show that identifying LiaX in invasive clinical isolates of E. faecium has potential to characterize strains in which the cell envelope stress response is activated, as better surrogate marker for DAP susceptibility. Such a test may be of value to overcome the major limitations of DAP MIC determination using current approaches.

MATERIALS AND METHODS

Bacterial strains and growth conditions

We used previously characterized reference E. faecium strains (20) as well as 86 E. faecium isolates (22). The reference and clinical strains used in this study are listed in Table S1. All clinical E. faecium isolates were recovered from the bloodstream of patients enrolled in the VENOUS, a prospective cohort study of US adults with enterococcal bloodstream infection (22). All reference and clinical strains have undergone whole genome sequencing as described previously (22) prior to the onset of this study. Enterococcal strains were grown on brain-heart infusion (BHI) agar or in BHI broth at 37°C with gentle agitation.

Transcriptional analysis of LiaR

The LiaR regulon of E. faecium R497 and its liaR null mutant derivative (R497ΔliaR) have been previously described (11, 13) using RNA-seq (32). qRT-PCR was performed to confirm RNA-seq results in relation to the level of expression of the liaXYZ cluster. Strains were grown to early exponential growth phase in BHI, and RNA extraction was performed using Purelink RNA Mini Kit (Ambion). Three biological replicates from each strain were performed. The samples were treated with Turbo DNAse kit (Ambion) to remove genomic DNA. The cDNA was obtained from ~1,000 ng of purified RNA using SuperScript II Reverse Transcriptase (Invitrogen). Evaluation of gene expression was conducted with 10 ng of cDNA using SsoAdvanced Universal SYBR Green Supermix (Bio-Rad) in CFX96 Touch TM Real-Time PCR Detection System (Bio-Rad). Relative expression ratios were calculated by normalizing to the housekeeping genes gyrB. The primer efficiency was calculated using a corrected calculation model described by Pfaffl (33). Primer efficiency was determined by LinRegPCR program for each reaction. Differences in gene expression between the pair of strains were calculated using the normalized expression for each gene. Primers used for qRT-PCR are listed in Table S4.

Bacterial mutagenesis and complementation

We generated non-polar deletions of liaR and liaX in strains of E. faecium. Deletion mutants of E. faecium TX1330RF were created using the p-chloro-phenylalanine (p-Chl-Phe) sensitivity counterselection system (PheS*) using pHOU1 plasmid, which confers resistance to gentamicin, as described previously (24). Briefly, upstream and downstream fragments of liaR and liaX were amplified by PCR using primers listed in Table S4. PCR products were cloned into pHOU1 using BamHI and EcoRI restriction sites. The plasmid constructs were electroporated into E. faecalis CK111 and subsequently delivered to E. faecium TX1330RF by conjugation. First-recombination integrants were selected on plates containing gentamicin 150 µg/mL and fusidic acid 20 µg/mL. Subsequently, these integrants were grown on p-Chl-Phe and selected based on gentamicin sensitivity. Gentamicin-susceptible colonies were screened for deletion of the desired gene by sequencing of the entire liaR or liaX open reading frames. Mutants were characterized by pulsed-field gel electrophoresis and antibiotic susceptibility testing.

For transcomplementation, the liaR gene was amplified by PCR (using primers in Table S4) using total DNA from E. faecium TX1330RF as template. The resulting fragment was cloned into pMSP3535 (34) under the control of the PnisA promoter which allows inducible expression in the presence of nisin. Recombinant pMSP3535 derivatives were purified on Luria-Bertani agar containing erythromycin (100 µg/mL). All inserts were confirmed by Sanger sequencing before introducing into E. faecium TX1330RF derivatives by electroporation. All E. faecium derivatives with pMSP3535 were maintained with erythromycin (20 µg/mL), and expression was achieved with nisin (50 ng/mL). Strains and plasmids used in mutagenesis are listed in Table S5.

Antimicrobial susceptibility testing

MICs for DAP for all E. faecium strains and isolates were performed via BMD and gradient strip. MICs for CRO were performed by BMD. For BMDs, serial twofold dilutions of antibiotics were prepared in a 96-well microtiter plate in Mueller-Hinton broth (BD) (supplemented with Ca₂Cl 50 mg/L for DAP BMDs). A 0.5 McFarland turbidity standard was prepared per strain, diluted to 5 × 10⁵ CFU/mL, and inoculated into microtiter wells. Broth microdilution assay preparations were conducted by two different researchers. Plates were read by three independent observers after incubation at 37°C for 24 hours. For gradient strip testing, a 0.5 McFarland turbidity standard was prepared per strain and inoculated on Mueller-Hinton agar. After bacterial solution absorption, DAP gradient Etest strips (bioMérieux, Inc., Durham, NC) were applied to the agar surface. After incubation at 37°C for 24 hours, MICs were read at the location where the elliptical zone of inhibition intersected the strip.

Expression and purification of LiaX from Enterococcus faecium

The gene encoding LiaX from R494 was cloned into a modified pET vector and transformed into E. coli BL21(DE3). LiaX was overexpressed by growing cells in Luria-Bertani broth containing 50 µg/mL kanamycin and inducing with 0.5 mM isopropyl β-D-1-thiogalactopyranoside for 20 hours at 16°C. Cells were pelleted and resuspended in buffer A (20 mM HEPES pH 7.4, 300 mM NaCl, 20 mM imidazole) with EDTA-free complete protease inhibitor tablets (Roche Diagnostics Corp., Indianapolis, IN) before undergoing lysis by sonication. Proteins were purified using a HisTrap FF crude column (GE Healthcare Life Sciences, Marlborough, MA). The protein was eluted with a continuous elution gradient of 20–500 mM imidazole. The LiaX fractions were pooled, and the His tag was subsequently removed by Tobacco Etch Virus (TEV) protease (35). The flowthrough was then dialyzed against 20 mM Tris-HCl pH 8.0 and purified over a Q-XL Sepharose column using a 0.1–1,000 mM NaCl gradient. The protein was further purified by size-exclusion chromatography, and the fractions containing LiaX were pooled and concentrated. Purification fractions and final pooled purified LiaX were run on 15% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), with purity assessed by ImageJ (National Institutes of Health) software.

Production and purification of polyclonal antibodies to LiaX

The purified LiaX protein was used to raise antisera in goats at Bethyl Laboratories (Montgomery, TX). The antisera subsequently underwent high-affinity antigen-based purification using CnBr-activated Sepharose coupled with purified LiaX protein (GeneMed Synthesis, Inc., San Antonio, TX). Antibodies were eluted with 0.1 M glycine-HCl, pH 2.5, and immediately neutralized after elution with 1 M Tris, pH 8.0.

ELISA antibody titer of goat polyclonal antibodies

To assess analytical sensitivity of the purified polyclonal goat antibodies to LiaX, an ELISA titer was conducted. Comparator goat pre-immune sera was used as a control to ensure absence of LiaX-binding antibodies prior to immunization with LiaX and to monitor for generation of nonspecific background signal due to antibody complexes in the absence of antigen. The ELISA was performed in triplicate in a 96-well microtiter plate with 10-ng purified LiaX per well and donkey anti-goat alkaline phosphatase-conjugated secondary antibody and chromogenic p-nitrophenyl phosphate (PNPP) substrate.

Western blotting for protein localization and analytical specificity

Strains were grown to mid-exponential phase. For localization assessment, cells were pelleted, and supernatant was aspirated. Supernatants were filtered through a 0.22 µM filter (Millipore, Sigma). In experiments examining the range of detection of the LiaX protein in supernatants, both aspirated and aspirated-then-filtered supernatants had a range of purified LiaX protein added in twofold dilutions: 2.5, 1.25, 0.625, 0.312, 0.156, 0.078, 0.039, and 0.019 ng. Protein was pelleted from the TCA supernatants at 15,000 rpm for 20 minutes at 4°C. Supernatant protein pellets were washed with 500-µL cold acetone twice, dried under vacuum for 60 seconds, and resuspended in 100 µL of phosphate-buffered saline (PBS) with phenylmethylsulfonyl fluoride protease inhibitor before mechanical lysis with glass beads at 4°C using a FastPrep 24 instrument. Cell debris and beads were spun down at 13,000 rpm for 20 minutes at 4°C. The cell lysates were collected, avoiding debris. Supernatant protein and cell lysates underwent bicinchoninic acid protein concentration assay in three biological replicates at 1:10 and 1:25 dilution for preparation of loading 10 µg of protein per SDS-PAGE well. Western blotting was performed on 8% SDS-PAGE gels and transferred to polyvinyldifluoride membranes using the iBlot 2 (Thermo Fisher Scientific). Membranes were blocked and probed with purified polyclonal goat antibodies to LiaX diluted 1:3,000, RNA polymerase β-subunit antibodies diluted 1:1,000 for protein loading control, and horseradish peroxidase-conjugated mouse anti-goat antibody diluted 1:3,000. Membranes were incubated with SuperSignal West Pico Chemiluminescent substrate (Thermo Scientific) and visualized using a ChemiDoc Imaging system (Bio-Rad). Relative intensities for the 58.6-kDa molecular weight band representing E. faecium LiaX were determined using the Gel Analyzer tool in Fiji (36) and normalized to the positive control purified E. faecium LiaX protein band.

The LiaX ELISA optimization and final protocol

The optimal antibody concentrations were determined using checkerboard titration. About 10⁷ CFUs of several E. faecium isolates with a wide range of LiaX production were coated to 96-well microtiter plate wells. Controls included 10-ng pure LiaX protein and wells without coated cells or other antigens. The dilutions of purified polyclonal goat antibodies to LiaX ranged from 1:100 to 1:4,000. The dilutions of donkey anti-goat alkaline phosphatase-conjugated secondary antibody ranged from 1:1,000 to 1:6,000. The optimal antibody dilutions were determined as the values that yielded the maximum A₄₀₅ values within the accurate detection range of the spectrophotometer without generation of non-specific background signal in negative control wells. The final LiaX ELISA was performed as follows: E. faecium isolates were grown to early mid-exponential growth phase and normalized to an A_600nm of 1.0 before cell pelleting and washing with PBS pH 7.4. The cells were then resuspended in 50 mM carbonate-bicarbonate buffer pH 9.6, and 100 µL (~10⁷ CFUs) was coated to wells in triplicate in 96-well high-binding microtiter plates for 16 hours at 4°C. Coated wells were washed three times with PBS, then blocked with 2% bovine serum albumin (BSA) 0.1% Tween 20 in PBS for 1 hour at room temperature. LiaX was detected with anti-LiaX purified polyclonal goat antibodies diluted 1:400 in 1% BSA and 0.05% Tween 20 in PBS and incubated for 2 hours at room temperature. After three washes with PBS, donkey anti-goat alkaline phosphatase secondary antibody was diluted 1:2,000 1% BSA, 0.05% Tween 20 in PBS, added to the wells, and incubated for 1 hour at room temperature. After three final washes with PBS, colorimetric substrate PNPP was added to the wells and incubated in the dark for 15 minutes at room temperature, and then, A₄₀₅ was read using a plate spectrophotometer. Isolates were tested in at least three separate assays with three technical replicates per assay.

Analysis of association of daptomycin MICs and the LiaX ELISA

A₄₀₅ values from the LiaX ELISA were plotted with DAP-R and DAP-S strains in an ROC curve. The cut-off was determined by maximizing the Youden index. Isolates were plotted according to MIC and positive versus negative LiaX ELISA.

LiaX and LiaR sequences and association with daptomycin MICs and the LiaX ELISA

The amino acid sequences of LiaX and LiaR (22) for each isolate were acquired. Amino acid sequences were aligned to the reference sequence of E. faecium DO using the alignment algorithm MUSCLE to determine amino acid substitutions (37). Amino acid sequences for LiaR and LiaX were then analyzed using the R statistical software with each amino acid position as a unique variable. A recursive partitioning model was used to compare each amino acid substitution as a variable in sequential decision trees to compare DAP MICs separately and A₄₀₅ of the LiaX ELISA for each isolate. To assess phylogeny of LiaX, maximum likelihood trees were constructed with LiaX protein alignments using RAxML (38) with the PROTGAMMAUTO model.

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/aac.01069-23.

Supplementary tables and figures. aac.01069-23-s0001.pdf.

Strains and plasmids, primers, LiaX ELISA assay precision, LiaX amino acid substitutions.

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.