Structure Activity Relationship of the Stem Peptide in Sortase A Mediated Ligation from Staphylococcus aureus

Abstract

The surfaces of most Gram‐positive bacterial cells, including that of Staphylococcus aureus (S. aureus), are heavily decorated with proteins that coordinate cellular interactions with the extracellular space. In S. aureus, sortase A is the principal enzyme responsible for covalently anchoring proteins, which display the sorting signal LPXTG, onto the peptidoglycan (PG) matrix. Considerable efforts have been made to understand the role of this signal peptide in the sortase‐mediated reaction. In contrast, much less is known about how the primary structure of the other substrate involved in the reaction (PG stem peptide) could impact sortase activity. To assess the sortase activity, a library of synthetic analogs of the stem peptide that mimic naturally existing variations found in the S. aureus PG primary sequence were evaluated. Using a combination of two unique assays, we showed that there is broad tolerability of substrate variations that are effectively processed by sortase A. While some of these stem peptide derivatives are naturally found in mature PG, they are not known to be present in the PG precursor, lipid II. These results suggest that sortase A could process both lipid II and mature PG as acyl‐acceptor strands that might reside near the membrane, which has not been previously described.

In Staphylococcus aureus , sortase A is the principal enzyme responsible for covalently anchoring proteins. Using a combination of two unique assays, we showed that there is broad tolerability of substrate variations that are effectively processed by sortase A. These results suggest that sortase A could process both lipid II and mature peptidoglycan near the membrane as acyl‐acceptor strands, which has not been previously described.

Introduction

Many pathogenic Gram‐positive organisms covalently anchor virulence proteins to the surface of their cell wall, a highly crosslinked meshwork known as peptidoglycan (PG). The PG is a thick outer layer that surrounds the cytoplasmic membrane in Gram‐positive organisms (Figure 1A). PG is typically comprised of the disaccharide N‐acetylglucosamine (GlcNAc) and N‐acetylmuramic acid (MurNAc) as a monomeric unit that is connected to form a polymeric scaffold. Attached to each MurNAc unit is a unique, short peptide (called the stem peptide) composed of up to five amino acids in length. The canonical pentapeptide sequence is typically L‐Ala‐iso‐D‐Glu‐L‐Lys (or meso‐diaminopimelic acid [m‐DAP])‐D‐Ala‐D‐Ala (Figure 1B). It is now well established that the primary sequence can have variation within a single bacterial cell but also across different organisms, especially at the third position. ^[1] A large fraction of bacterial species have an m‐DAP residue at the third position ^[2] and most of the remaining have L–Lys in this position. A common observation for organisms with L‐Lys in the third position, is that additional amino acids are often connected to amino group on the L‐Lys sidechain, and these additional amino acids are collectively known as cross‐bridging amino acids. Cross‐bridging amino acids can include ‐ but are not limited to ‐ L‐Ala, D‐Ser, D‐iAsx (where Asx indicates aspartic acid or asparagine), and Gly.[ ² , ³ ]

Figure 1.

(A) Schematic representation of the surface of S. aureus. The enzyme SrtA covalently anchors proteins onto the peptidoglycan layer, which are imbedded within and displayed onto the surface of these bacteria. (B) Canonical monomeric unit of peptidoglycan for S. aureus, which includes the disaccharide backbone and the stem peptide. Site of acyl‐acceptor nucleophile is shown in red.

The location of the PG matrix provides an opportunity for Gram‐positive organisms to stage proteins that support functions that are important for survival and proliferation. A primary mode of protein display involves the covalent anchoring of proteins within the PG matrix, a reaction carried out by sortases. Staphylococcus aureus (S. aureus) is a prominent example of a human pathogen that utilizes sortase‐mediated anchoring to acquire iron, promote adhesion, and subvert host immune response. ^[4] More specifically, immune evasion occurs by the display of protein A, which binds the Fc region of immunoglobin G (IgG) during infection to reduce recognition by the host immune system. ^[5] Sortase enzymes are transpeptidases that recognize a specific sorting sequence present in the protein for eventual anchoring within the PG. ^[6] In S. aureus, the sorting sequence for processing by sortase A (SrtA) is LPXTG, whereby X is any amino acid. ^[7] The cysteine residue in the active site of SrtA clips the peptide bond between T and G of the sorting sequence, thus forming a thioacyl intermediate between sortase and the threonine of the LPXT‐containing protein. ^[8] This thioacyl intermediate undergoes a nucleophilic attack by the N‐terminal amino group, which is found on the pentaglycine (G₅) cross‐bridge of S. aureus PG (Figure 2). ^[9] A new amide bond is created between the protein and the PG pentaglycine cross‐bridge, allowing for covalent display of protein virulence factors on the bacterial cell surface. ^[10]

Figure 2.

Similarity in the two reactions that use a common acyl‐acceptor strand. In S. aureus, cell wall crosslinking of neighboring strands is performed by PBPs and Ldts (left). A similar reaction pathway for the second step of the reaction is observed for SrtA, with the primary difference being the acyl‐donor strand.

Due to the fact that SrtA aids S. aureus in host‐colonization and altered host‐immune response, it is considered a promising anti‐virulence target. The deletion of SrtA in S. aureus disrupts the establishment of infection in mouse models. ^[11] These results have spurred considerable efforts to discover effective SrtA inhibitors. ^[12] Additionally, the transpeptidase activity of sortase has been harnessed for use as a general ligation tool outside the context of bacterial cell surfaces. This has been demonstrated by extensive work whereby various synthetic nucleophiles have been used, as well as a multitude of LPXTG‐containing scaffolds, in place of the natural substrates found within the bacterial cell, to ligate molecules of interest. ^[13] Interestingly, the transpeptidase reaction performed by sortase has functional similarity to the transpeptidases responsible for PG crosslinking, carried out by either penicillin binding proteins (PBPs) or L,D‐transpeptidases (Ldts). These transpeptidase enzymes use two neighboring strands of PG to generate 4–3 (between the 4^th position of the donor strand and 3^rd position of the acceptor strand) or 3–3 (between the 3^rd position on both strands) crosslinks. ^[14] The primary difference between these two types of transpeptidases occurs in the first half of the reaction. Whereas PG crosslinking enzymes process a stem peptide as the acyl‐donor strand, SrtA will instead process the LPXTG (Figure 2). The second half of the reaction ‐ the nucleophilic attack from amino group on the G₅ crossbridge ‐ potentially utilizes the same substrate.

Given the essentiality of PG crosslinking, as demonstrated by the number and diversity of small molecule antibiotics that disrupt this pathway, it is possible that dysregulated processing of the same substrate (acyl‐acceptor strand) could lead to loss of cell wall integrity. More specifically, we wondered whether the regulation of acyl‐acceptor utilization could be inherently dictated by the primary sequence of the stem peptide. It is well established that the stem peptide in S. aureus, and for most bacteria whose PG have been analyzed, can vary in length (penta‐, tetra‐, and tri‐peptides all retain the acyl‐accepting crossbridge) ^[15] and amidation states of D‐iGlu. ^[16] To investigate this, we systematically measured sortase activity of SrtA in vitro with variations of possible acyl‐acceptor stem peptides endogenously found within the PG of S. aureus. Interestingly, there was not a strong preference within the stem peptide primary sequence variations. It had been described that the acyl‐donor strand for SrtA is lipid II. ^[10a] Our results suggest that it may be possible that the acyl‐donor strand comes from a more mature PG scaffold residing near the membrane. This could indicate that sortase‐mediated transpeptidase activity can be decoupled from lipid II pools.

Results and Discussion

We first set out to measure the activity of sortase A using gel fluorescence analysis. ^[17] We synthesized (using standard solid phase peptide coupling procedures) a truncated substrate of the acyl‐acceptor strand GGGK(5,6‐carboxytetramethylrhodamine, TAMRA; GGGK(tmr)) ^[18] whereby a red‐fluorophore is connected to the sidechain of lysine. Additionally, we expressed and purified a green fluorescent protein (GFP) containing a C‐terminal LPETG sorting signal to serve as the acyl‐donor fragment. Reactions were carried out using standard conditions and following this incubation period, the reaction contents were boiled in sodium dodecyl sulfate (SDS) to denature the GFP. Accordingly, the fluorescence observed is expected to come solely from a productive transpeptidation of the TAMRA‐containing strand onto the larger molecular weight of the modified GFP. An expected product band of ∼29 kDa was anticipated for GFP‐LPETGGGK(tmr) products and was observed by gel analysis (Figure 3A). Reactions were also set up in the absence of each reaction constituent (e. g., no SrtA, no GFP, or no GGGK(tmr)) as negative controls. Using this set up, a time point assay was also performed, where the reaction with all elements was quenched at 1, 2, 4, 8 h ^[19] and displayed an increase in fluorescence as time progressed (Figure S1).

Figure 3.

(A) Gel fluorescence analysis of GGGK(tmr) in the various stated conditions. Reaction was performed at 37 °C and allowed to reaction for 8 h, then loaded on an SDS‐PAGE. (B) Gel fluorescence analysis of GFP‐LPETG and SrtA in the presence of tmrP^Egly₁ , tmrP^Egly₃ , or tmrP^Egly₅ . Reaction was performed at 37 °C and allowed to reaction for 2 h, then loaded on an SDS‐PAGE.

With reaction parameters optimized, we assembled a panel of stem peptide mimetic probes that would allow us to interrogate the necessity of naturally occurring peptide lengths or modifications for SrtA activity. Our group, and others, previously showed that synthetic acyl‐acceptor strands modified with an N‐terminal fluorophore are well tolerated by cell wall crosslinking enzymes (PBPs and Ldts). ^[20] We envisioned that a similar tolerability would enable the structure‐activity relationship of SrtA. As wildtype S. aureus displays pentaglycine cross‐bridge (G₅) appended to its 3^rd position lysine sidechain, we first sought to determine the importance of the glycine cross‐bridge length. Collectively, biosynthesis of this cross‐bridge is performed by enzymes in the fem gene family (femA, femB, femAB, femAX).[ ³ , ²¹ ] It had been shown that non‐pentaglycyl cross‐bridges have the ability to ligate surface proteins to the PG, although the natural nucleophilic cross‐bridge is understood to be preferred.[ ⁹ , ²² ] We synthesized a penta‐stem peptide with a TAMRA‐modified N‐terminus and varied the glycine bridge length as 5 (tmrP^Egly₅ ), 3 (tmrP^Egly₃ ), or 1 (tmrP^Egly₁ ) glycine(s) (Figure 3). Reactions were set up with each of the probes, SrtA, and GFP‐LPETG and analyzed by in‐gel fluorescence. Interestingly, there appeared to be a slight increase in transpeptidation with the native glycine cross‐bridge at 2 h. A distinguishably brighter signal for tmrP^Egly₅ was observed over both tmrP^Egly₃ and tmrP^Egly₁ (Figure 3B). The differences in fluorescence intensity became smaller with longer incubation time, which may indicate that subtle differences in SrtA activity may only be observable in early time points prior to reaching reaction equilibrium (Figure S2).

While the in‐gel fluorescence method could qualitatively discern large changes in transpeptidation preferences, we realized that the structure‐activity relationship in the acyl‐acceptor chain may be subtle. Therefore, it was important to develop an assay with a higher level of signal‐to‐noise ratio and sensitivity. We reasoned that it is possible load the sorting signal onto a flow cytometry‐compatible bead and co‐incubate with SrtA and fluorescent stem peptide analogs linked to a fluorophore for analysis by flow cytometry (Figure 4). Upon transpeptidation of sortase, the fluorescent handle would be covalently attached to the bead, which could be readily and efficiently analyzed by flow cytometry. The carboxyl beads were first modified with an amino polyethylene glycol (PEG) spacer functionalized with an azido group (H₂N‐PEG₂₃‐N₃) to provide better display of the sorting signal away from the surface of the bead. ^[13d] The azido group on the bead was then reacted with K(DBCO)LPMTG to install the SrtA substrate via strain‐promoted click chemistry. ^[23] To ensure that H₂N‐PEG₂₃‐N₃ was successfully coupled to the beads, we initially treated some of the beads with DBCO−Fl, washed away excess DBCO−Fl, and monitored the fluorescence of the beads via the use of flow cytometry. An observable increase in fluorescence over beads without H₂N‐PEG₂₃‐N₃ was recorded (Figure S3), confirming that the beads were displaying N₃ on the surface.

Figure 4.

Schematic representation of the sortase‐mediated ligation of the fluorescently‐linked stem peptide analog onto a bead. The beads are analyzed by flow cytometry and the fluorescence level is reflective of the sortase‐related ligation.

With the KLPMTG linked beads in hand, we turned our attention to quantitatively analyze SrtA reactions. For all bead‐based assays, the stem peptides were modified with fluorescein (fl) instead of TAMRA due to better compatibility for the flow cytometer. Initially, the amidation of D‐iGlu was analyzed in the context of a full length pentapeptide in the molecules flP^Qgly₅ and flP^Egly₅ (Figure 5A). Amidation at the carboxyl of the iso‐D‐glutamic acid residue (which results in iso‐D‐glutamine) is a PG modification catalyzed by the MurT and GatD enzyme complex. ^[24] There has been some speculation about the role of amidation for PG assembly. To this end, prior in vitro studies using enzymes from Streptococcus pneumoniae, S. aureus, Mycobacteria, and Enterococcus faecalis suggested that PG crosslinking by PBPs is disrupted in the absence of amidation.[ ^20a , ²⁵ ] In fact, the MurT/GatD complex is essential in a number of human pathogens, making it a potential drug target.[ ^25b , ²⁶ ] As expected, due to SrtA‐mediated ligation of a fluorescently tagged stem peptide to the sorting signal‐containing beads, only when sortase was present was a fluorescent readout observed (Figure 5B). Interestingly, there was only a small difference in fluorescence levels between flP^Qgly₅ and flP^Egly₅ in this assay, which suggests that amidation of iso‐D‐Glu is not playing a large role in controlling acyl‐acceptor strand processing when 5 glycines are present on the cross bridge.

Figure 5.

(A) Chemical structures of flP^Qgly₅ and flP^Egly₅ . (B) Flow cytometry analysis of beads modified with the SrtA sorting signal, then treated with stem peptide analogs in the presence or absence of SrtA. Data are represented as mean +/− SD (n=3). P‐values were determined by a two‐tailed t‐test (* denotes a p‐value < 0.05, ** < 0.01, ***<0.001, ns = not significant).

Owing to the fact that it is possible that SrtA is operating on stem peptides of various lengths, we turned our attention to shorter versions of the stem peptide (tetrapeptides, tripeptides) that are known to be highly abundant within the PG scaffold of S. aureus. ^[27] While tetra‐ and tripeptides can act as substrates in transpeptidase‐catalyzed reactions for PG crosslinking, we wanted to examine the extent to which they participate in sortase A‐mediated ligation reactions. As such, we synthesized a series of trimeric and tetrameric stem peptide analogs to examine the extent to which sortase utilizes them as substrates using the bead assay (Figure 6A). Our data showed that there is not a large effect in terms of both amidation and stem peptide length. From these results, it appears that SrtA does not have a strong preference for a specific length of the stem peptide or amidation of D‐iGlu. These results provide evidence that it may be possible for SrtA (and other related sortases) to covalently link proteins onto the mature PG that might reside near the membrane, as well as the nascent PG unit. To test this concept, we isolated sacculi from S. aureus and incubated it with SrtA and a fluorescein‐tagged LPMTG sorting signal (Fl‐LPMTG). ^[12a] Similar to the bead assay, we expected that processing by SrtA would result in the covalent linking of LPMTG onto the sacculi, which was analyzed by flow cytometry (Figure 6C). The sacculi are devoid of lipid II, and, instead, only will have mature PG. Our results showed that efficient transpeptidation is observed in the absence of lipid II (Figure 6D). Inhibition of SrtA with methanethiosulfonate (MTSET), a covalent inhibitor of sortase ^[28] led to a significant decrease in fluorescence levels. Moreover, when the acyl‐donor sites are chemically blocked by acetylation, there is a significant drop in fluorescence levels confirming the necessity of free amino groups in the cross‐bridge for acylation. There were increasing levels of fluorescence associated with the sacculi with increasing concentrations of Fl‐LPMTG (Figure S4). Moreover, the solution phase reaction showed that the product was formed as expected (Figure S5). Together, these results are consistent with the prospect that processing by SrtA may be possible in various PG fragments and may not be restricted to lipid II. In support of this, it was recently shown that incubation of bacterial cells with LPXTG modified probes led to incorporation within the mature PG, not lipid II. ^[29]

Figure 6.

(A) Chemical structure of tripeptide and tetrapeptide analogs of S. aureus PG. Each has a fluorescein linked to its N‐terminus for quantification. (B) Flow cytometry analysis of beads modified with the SrtA sorting signal, then treated with various stem peptide analogs or absence (DMSO) in the presence of SrtA. Data are represented as mean +/− SD (n=3). (C) Schematic representation of the SaccuFlow analysis of the SrtA processing of the sorting signal modified with a fluorescent handle. (D) Flow cytometry analysis of sacculi incubated with the stated conditions in the presence of Fl‐LPMTG. Data are represented as mean +/− SD (n=3).

Conclusion

We have assembled a systematic analysis of the possible stem peptides from S. aureus that can act as acyl‐acceptor strands. These analogs probed the length of the cross‐bridge, the amidation state of D‐iGlu, and the length of the stem peptide backbone. To quantitatively measure SrtA activity, we developed a flow‐based bead assay that is robust and has the potential to be compatible with high‐throughput analysis. Our results showed that SrtA activity was mostly independent of the individual structural modifications we selected. These results could indicate that SrtA can, potentially, anchor proteins onto mature PG and does not necessarily exclusively utilize lipid II as the acyl‐acceptor strand, as it had been described previously. To the best of our knowledge, this is the first analysis of stem peptide variation in the activity of SrtA. We anticipate that a similar platform can be applied to other sortases or enzyme classes (e. g., Braun protein from Escherichia coli) to better understand how the primary structure of the PG could potentially modulate its processing by cell wall‐linked enzymes.

Conflict of interest

The authors declare no conflict of interest.

Supporting information

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re‐organized for online delivery, but are not copy‐edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

Supporting Information

A. J. Apostolos, J. J. Kelly, G. M. Ongwae, M. M. Pires, ChemBioChem 2022, 23, e202200412.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.