Chemical tools for selective activity profiling of bacterial penicillin-binding proteins

Abstract

Penicillin-binding proteins (PBPs) are membrane-associated proteins involved in the biosynthesis of peptidoglycan (PG), the main component of bacterial cell walls. These proteins were discovered and named for their affinity to bind the β-lactam antibiotic penicillin. The importance of the PBPs has long been appreciated; however, specific roles of individual family members in each bacterial strain, as well as their protein-protein interactions, are yet to be understood. The apparent functional redundancy of the 4–18 PBPs that most eubacteria possess makes determination of their individual roles difficult. Existing techniques to study PBPs are not ideal because they do not directly visualize protein activity and can suffer from artifacts and perturbations of native PBP function. Therefore, development of new methods for studying the roles of individual PBPs in cell wall synthesis is required. We recently generated a library of fluorescent chemical probes containing a β-lactone scaffold that specifically targets the PBPs, enabling the visualization of their catalytic activity. Herein, we describe a general protocol to label and detect the activity of individual PBPs in Streptococcus pneumoniae using our fluorescent β-lactone probes.

1. Introduction

Peptidoglycan (PG) is a heteroglycan polymer crosslinked by short peptide chains branched off of the glycan backbone that makes up the primary component of the bacterial cell wall. PG is essential for bacterial cell growth and survival, and as such presents an ideal target for antibiotics, as the cell wall and the concomitant machinery are exclusively present in prokaryotic organisms. Penicillin-binding proteins (PBPs) are a family of bacterial enzymes responsible for the synthesis of PG via the polymerization of the glycan chain (transglycosylation) and the cross-linking of peptide chains (transpeptidation). β-Lactam antibiotics such as penicillin inhibit the transpeptidation reaction by covalently binding to the PBPs, resulting in cell lysis, and as such are some of the most effective treatments for bacterial infection. Despite the importance of the PBPs in bacterial growth and the clinical importance of cell wall-targeting antibiotics, little is known about the functions and regulation of individual PBP isoforms within the same organism. Most bacteria contain multiple PBPs that are highly structurally homologous and functionally redundant. The inability to elucidate the catalytic activity, localization, and regulation of individual PBPs represents a critical gap in knowledge in the battle against antimicrobial resistance.

Current strategies to study bacterial cell wall synthesis and remodeling have relied extensively upon the β-lactam scaffold since it was determined that penicillin V is a global inhibitor of PBPs (Böttcher & Sieber, 2012; Falconer, Czarny, & Brown, 2011). A common strategy to study PBP catalytic activity is to use radiolabeled or fluorescently-tagged penicillin analogs, followed by gel separation and visualization of the tagged proteins (Ghuysen, 1977; Lakaye et al., 1994; Spratt & Pardee, 1975; Waxman & Strominger, 1983; Zhao, Meier, Kahl, Gee, & Blaszczak, 1999). However, this strategy cannot report on individual PBP catalytic activity, as most penicillin-based compounds inhibit all PBPs within a given organism. Fluorescent fusion proteins of the PBPs have been utilized, but these artificial fusions can interfere with native protein abundance, localization, or activity (Margolin, 2012; Swulius & Jensen, 2012). Furthermore, such fusion tags only report on protein localization, not activity.

Fluorescent small molecules offer spatiotemporal resolution while eliminating the need for genetic manipulation. For example, newly synthesized PG can be visualized with fluorescent vancomycin (Van-FL) and ramoplanin, which bind to PG precursors in Gram-positive bacteria (Daniel & Errington, 2003; Tiyanont et al., 2006). Other strategies have used fluorescent D-amino acid analogs (FDAAs) that are incorporated into the PG by PG synthesis machinery, enabling visualization of nascent peptidoglycan molecules (Fura, Kearns, & Pires, 2015; Kuru et al., 2012; Lebar et al., 2014; Pidgeon et al., 2015; Siegrist et al., 2013).

In order to tease apart the various roles of the PBPs and regulation of PG synthesis and remodeling, β-lactam-based probes have been synthesized for in vitro labeling, primarily targeting the PBPs in addition to other bacterial proteins (Dargis & Malouin, 1994; Staub & Sieber, 2008; Zhao et al., 1999). The Carlson lab reported that the antibiotic cephalosporin C could be modified to produce a chemical probe that selectively labeled a subset of PBPs in S. pneumoniae and Bacillus subtilis (Kocaoglu et al., 2012). In order to identify selective probes for individual PBP isoforms, a library of 20 commercially available β-lactams was screened against S. pneumoniae, B. subtilis and Escherichia coli (Kocaoglu & Carlson, 2015; Kocaoglu, Tsui, Winkler, & Carlson, 2015; Sharifzadeh, Dempwolff, Kearns, & Carlson, 2020). Of the screened compounds, several possessed selectivity for one or several PBPs at different concentrations of the β-lactam. However, many PBPs were not inhibited by available β-lactams, indicating a need to further explore chemical space. To that end, the Carlson lab has identified a β-lactone scaffold as a promising foundation from which selective PBP inhibitors can be developed and has applied these reagents to the selective imaging of essential PBPs, PBP2x and PBP2b, in S. pneumoniae.

2. Activity-based profiling of penicillin-binding proteins

Activity-based protein profiling (ABPP) methods have been utilized for the last two decades as a means to investigate complex proteomes in all three domains of life. Extensive use in eukaryotes has provided a more fundamental understanding of the biological processes involved in numerous diseases, which has played a key role in drug discovery campaigns. However, the use of ABPP in prokaryotes has lagged behind and has only gained more attention over the last decade. The alarming rise in antibiotic resistance has called for a better understanding of bacterial physiology and pathogenicity at a foundational level. ABPP provides a means by which new, clinically relevant drug targets may be identified through gaining insight into key biological processes. Several reviews are available on the use of ABPP, including their utility in prokaryotic systems (Cravatt, Wright, & Kozarich, 2008; Keller, Babin, Lakemeyer, & Bogyo, 2019; Krysiak et al., 2017; Sharifzadeh, Shirley, & Carlson et al., 2019). This chapter elaborates on the activity-based probes (ABPs) that we have applied to selectively profile individual members of the PBP enzyme family. Here, the methodology employed to design and apply a novel class of PBP ABPs containing a β-lactone scaffold is discussed in detail (Sharifzadeh et al., 2017). Our studies focused on S. pneumoniae, an ovoid-shaped Gram-positive bacteria that causes serious diseases such as pneumonia, bacteremia, and meningitis, and is the major cause of worldwide childhood mortality from infectious disease (Henriques-Normark & Tuomanen, 2013). In addition to being clinically significant, S. pneumoniae is an important model of PG biosynthesis as it possesses a relatively simple PBP profile, with three class A PBPs (PBP1a, PBP1b, and PBP2a), two class B PBPs (PBP2x and PBP2b), and one class C PBP (PBP3 or DacA; D,D-carboxypeptidase).

2.1. Activity-based probes: What they are and how they work?

ABPs represent any active-site-directed chemical probe that reacts with the enzyme in a mechanism-based manner, covalently attaches to a catalytic residue, and is retained in the active site (Heal, Dang, & Tate, 2011). As such, ABPP has emerged as a powerful approach to monitor the activity of proteins in complex proteomes. This strategy enables exclusive visualization of the catalytically active form of an enzyme. ABPs are mechanism-based irreversible inhibitors consisting of an active warhead, which reacts covalently with a nucleophilic active-site residue, a binding or spacer group that provides spacing between the active warhead and the reporter, which is a moiety that serves as a tag for identification and purification purposes (Fig. 1A). The active warhead is typically an electrophile that selectively reacts with an active-site nucleophile and is retained on-site due to the formation of a covalent bond. Since only catalytically competent enzymes can undergo this reaction, active isoforms are selectively visualized. Upon enzyme binding, the reporter tag of ABPs enables direct visualization of the probe (gel-based or imaging) or enrichment and protein identification (mass spectrometry) depending on the type of reporter used (Sadaghiani, Verhelst, & Bogyo, 2007). Initial ABPP studies took advantage of direct radiochemical detection through incorporation of ¹²⁵I, which is not applicable in large scale analysis due to limitations imposed on radioactive material (Bogyo, Verhelst, Bellingard-Dubouchaud, Toba, & Greenbaum, 2000). More recently, fluorescent and affinity tags were incorporated and have been utilized to detect enzymatic labeling of ABPs (Patricelli, Giang, Stamp, & Burbaum, 2001). Biotinylated ABPs have been developed to enable target enrichment, a more time-consuming detection strategy. Fluorescent ABPs are the most direct method for in-gel detection of activity with high sensitivity and throughput.

Fig. 1.

(A) General structure of an ABP. An electrophilic group is used most commonly for the reactive group, which is linked to a biotin or fluorophore molecule by an unreactive linker. (B) General workflow of ABPP. An ABP is applied to a sample of interest, which can then be processed in a multitude of ways, depending on the application. The end-result is a labeled proteome that may be visualized via gel-based analysis or mass spectrometry (MS). Analysis of the enriched proteome could provide additional information about the enriched proteins such as identity, quantity and the site of labeling. Adopted from Sharifzadeh, S., Shirley, J. D., & Carlson, E. E. (2019). Activity-based protein profiling methods to study bacteria: The power of small-molecule electrophiles. Current Topics in Microbiology and Immunology, 420, 23–48. doi:10.1007/82_2018_135.

The irreversible, covalent nature of the ABPs’ bond with their targets has made them amenable for diverse applications including live-cell imaging, target enrichment and direct biochemical analysis of the tagged proteins (Fig. 1B). Moreover, assessment of the selectivity of enzyme inhibitors is another potential application for activity-based proteomics. In bacterial systems, ABPs have been applied for various applications such as target discovery, the study of microbial pathogenesis and host-pathogen interactions, live-cell imaging, and inhibitor discovery (Böttcher & Sieber, 2008; Heal et al., 2011; Sadler & Wright, 2015; Sharifzadeh et al., 2017; Willems, Overkleeft, & van Kasteren, 2014).

2.2. PBP-selective ABPs

The PBPs are PG synthase enzymes that were discovered as the targets of the β-lactam antibiotic, penicillin. Aside from their therapeutic application, β-lactams have been used as chemical probes to study the PBPs (Chambers, Sachdeva, & Hackbarth, 1994; Curtis, Orr, Ross, & Boulton, 1979; Kocaoglu et al., 2012; Staub & Sieber, 2008; Williamson, Hakenbeck, & Tomasz, 1980; Zhao et al., 1999). β-Lactams structurally resemble the terminal D-Ala-D-Ala moiety of PG stem peptides, and hence the transpeptidation (TP) domain of the PBPs recognizes them as their natural substrate. During the TP step, a conserved serine residue in the active site attacks the β-lactam ring in a mechanism-based manner, leading to formation of an inert acyl-enzyme complex that inhibits biosynthesis of the cell wall. Given the chemical stability of this acyl-PBP intermediate, β-lactams have long been used as chemical probes to study the PBPs. While radioactive penicillins were commonly applied in this regard for several decades (Chambers et al., 1994; Preston, Wu, Blaszczak, Seitz, & Halligan, 1990; Spratt & Pardee, 1975), fluorescent analogs were more recently developed and quickly replaced their radioactive counterparts in profiling the activity of the PBPs (Kocaoglu et al., 2012; Zhao et al., 1999). Bocillin-FL (Boc-FL), comprised of penicillin V as the active warhead and BODIPY FL as a fluorescent reporter group, is commercially available. Showing global affinity for the entire PBP content in a given microorganism, Boc-FL has been exploited to monitor TP activity in pure protein samples, as well as cell lysates and whole cells in multiple microbes (Kocaoglu & Carlson, 2015; Kocaoglu et al., 2015; Stone, Butler, Phetsang, Cooper, & Blaskovich, 2018; Zhao et al., 1999). In order to obtain selectivity for an individual PBP or subset of PBPs, cephalosporin C molecules tagged with different fluorophores were developed by our lab, which targeted a subset of PBPs in B. subtilis and S. pneumoniae (Kocaoglu et al., 2012).

2.2.1. Design rationale

The β-lactone (2-oxetanone) scaffold was inspired by its presence in a number of natural products showing antibacterial activity, as well as structural mimicry of the D-Ala-D-Ala portion of the stem peptide (Fig. 2A). The strained β-lactone ring is an important pharmacophore which is present in diverse natural products and biologically active scaffolds, some of which are known to possess antibiotic activity including orlistat, obafluorin, SQ-26,517, and hymeglusin (Aldridge, Giles, & Turner, 1971; Parker, Rathnum, & Liu, 1982; Ravindran et al., 2013; Tymiak, Culver, Malley, & Gougoutas, 1985; Wells et al., 1982). β-Lactone compounds have been shown to covalently modify enzymes, including serine hydrolases (Kim et al., 2002). As such, β-lactone drugs have application in the treatment of obesity, diabetes and cancer (Basha & Sankaranarayanan, 2016; Lawasut et al., 2012; Venukadasula, Chegondi, Maitra, & Hanson, 2010). The most well-known β-lactone, tetrahydrolipstatin (THL; Orlistat®), is a long-term anti-obesity drug that functions by irreversibly inhibiting a lipase (Venukadasula et al., 2010).

Fig. 2.

β-Lactone-based probe library. (A) The design of the β-lactone probe library was inspired by the structural similarity between the terminal D-Ala-D-Ala moiety of PG as the natural substrate of PBPs, the β-lactam core in antibiotics and the β-lactone scaffold present in antibacterial natural products such as SQ-26,517. (B) A library of probes was designed containing the electrophilic threonine β-lactone ring as the warhead and different fluorophore groups were implemented as the reporter. Probes with altered stereochemistry of the β-lactone were included to study the importance of that feature. Different amino acids carrying a variety of side chains were introduced into the probe structure to expand the PBP binding spectrum. (C) Chemical structures of the synthesized probes.

Given the potential of the β-lactone scaffold to interact with active-site serine residues, including those in β-lactamases and a carboxypeptidase (Böttcher & Sieber, 2008; Compton, Schmitz, Sauer, & Sello, 2013; Wells, Trejo, Principe, & Sykes, 1984; Zeiler et al., 2011), and the structural similarity of several of these compounds to the β-lactam antibiotics, we postulated that this scaffold could be optimized to create PBP-selective probes. Additionally, our assessment of the natural product SQ-26,517 indicated that it is a weak inhibitor of PBP2a in S. pneumoniae (Sharifzadeh et al., 2017). We designed our molecules with an amide group proximal to the electrophilic carbonyl and the ring substituents in the cis conformation to mimic SQ-26,517 and the β-lactam antibiotics. Structural diversity was produced through insertion of different amino acids, with small side chains such L-alanine and D-alanine (to mimic the side chains present in the natural substrate of PBPs), as well as hydrophobic and aromatic groups such as L-valine, L- and D-phenylalanine (to mimic the side chain of penicillin G), L-tyrosine and L-tryptophan. Three fluorophores, fluorescein, BODIPY FL and tetramethylrhodamine (TAMRA) were introduced as reporter groups (Fig. 2C).

The probe library contained 24 molecules: six that possess a lactone coupled directly to a fluorophore [Fig. 2C; 2–4; TAMRA (T), BODIPY FL (B) or fluorescein (FL)], six compounds with substrate-like sidechains (5 and 6), eight compounds with sidechains intended to mimic groups found in β-lactam antibiotics (7–9), two functionalized with glycine to assess the role of the amino acid side chain in protein labeling (10), and two with alternative hydrophobic side chains (L-Val and L-Trp, 11 and 12).

2.2.2. Results

Assessment of both cis-functionalized analogs, (2R,3S)-β-lactone (2) and (2S,3R)-β-lactone (3), demonstrated that the stereochemistry of the substituents on the lactone ring is crucial for labeling of the PBPs (Fig. 3). The compounds displaying these groups in the same orientation as is found in β-lactams, (2R,3S)-β-lactone (2FL and 2T), labeled most PBPs (PBP1a, PBP1b, PBP2x and PBP2a) in S. pneumoniae, while the opposite lactone configuration, (2S,3R)-β-lactone (3T), failed to label any PBPs (Fig. 3). We also found that lactone labeling of the PBPs is competitive with penicillin, as illustrated by the results of penicillin V pretreatment studies, where cells pretreated by penicillin V prior to labeling by β-lactone probes showed no labeling (data not shown). Moreover, treatment of β-lactone-labeled cells by penicillin V had no effect on PBP labeling (data not shown). These findings suggest that the β-lactone probes are active-site directed similar to the β-lactam antibiotics, and that the resulting acylated protein complex is stable, as it is not displaced by subsequent incubation with penicillin V (Sharifzadeh et al., 2017).

Fig. 3.

Gel-based analysis of probe labeling of S. pneumoniae PBPs to examine ring geometry (2 and 3), side chains that act as substrate mimetics (5 and 6), side chains that mimic β-lactam antibiotics (7–9), and side chains of different sizes and hydrophobicity (10 12). Results of representative probes from each group are shown here, while the full labeling profile can be found in Sharifzadeh et al. (2017).

Overall, we found that PBP1b is labeled by all the tested β-lactone probes, while several compounds co-selectively labeled PBP1b and PBP2x. This may indicate that these proteins are tolerant of side chains with diverse size and configuration. Intriguingly, PBP2x is a common target of the β-lactams in S. pneumoniae and is often co-inhibited with PBP3 (Kocaoglu et al., 2015). PBP3 was not labeled by any of the β-lactone probes in our study. In contrast, PBP1b is among the less frequently targeted PBPs by the β-lactam antibiotics (Kocaoglu et al., 2015), indicating that our probe library may be accessing a different region of binding space within the PBP active sites. PBP2b, which is the least inhibited PBP by conventional β-lactam antibiotics, was labeled by three probe scaffolds (5T, 5FL, 8T, 8FL, 10T). PBP1a and PBP2a labeling was only observed with compounds displaying a small side chain (Fig. 3).

PBP2x is an essential class A PBP in S. pneumoniae involved in septal PG synthesis (Berg, Stamsas, Straume, & Havarstein, 2013; Land et al., 2013; Tsui et al., 2014). Although we did not identify a probe that exclusively labeled PBP2x, we did uncover several compounds that co-labeled PBP2x and PBP1b, the latter of which can be deleted to yield a strain that is phenotypically indistinguishable from the wild-type organism (Land & Winkler, 2011). Thus, the Δpbp1b mutant (E193) provides the ideal platform in which to image solely the activity and localization of PBP2x. We utilized the 7FL probe in imaging studies to localize PBP2x during the course of cell division in S. pneumoniae E193 cells (Figs. 4–6).

Fig. 4.

PBP profiles of S. pneumoniae WT (IU1945), Δpbp1b (E193) and Δpbp1a (E177) cells that are labeled with Boc-FL, 7FL and 8T probes to resolve PBP1a and PBP1b bands. All probes were used at 5μg/mL. Dual labeling with (2R,3S)-β-lactone-L-Phe-fluorescein (7FL) followed by (2R,3S)-β-lactone-D-Phe-TAMRA (8T; dual labeling indicated by arrow above the right-hand gel) enables separate visualization of the activity of PBP2x (green) and PBP2b (red). Adapted from Sharifzadeh, S., Boersma, M. J., Kocaoglu, O., Shokri, A., Brown, C. L., Shirley, J. D., et al. (2017). Novel electrophilic scaffold for imaging of essential penicillin-binding proteins in Streptococcus pneumoniae. ACS Chemical Biology, 12(11), 2849–2857. https://doi.org/10.1021/acschembio.7b00614.

Fig. 6.

Inhibition of PBP2x by methicillin. (A) Lactone labeling of PBP2x can be prevented by pretreating cells with methicillin. Treatment was carried out using 0.1μg/mL methicillin for 30min prior to labeling with Boc-FL or 7FL (at 1 or 5μg/mL, as indicated by the numbers in the parentheses). Boc-FL was used at 5μg/mL as a control. (B) 7FL labeling of PBP2x in wild-type S. pneumoniae (IU1945) and Δpbp1b (E193) cells and a PBP2x-inhibiting concentration of methicillin. Cells were grown and pretreated with methicillin (0.1μg/mL), or PBS as a control, for 20min, followed by labeling with 7FL. The top row consists of different cells, with the white cartoon showing the orientation of each. The bottom two rows are pairs of images of the same cell, with rotations around the indicated axis by the white arrow. The lack of PBP1b does not prevent 7FL labeling, but methicillin-treated cells have been elongated and do not possess septal labeling, which is indicated by the empty arrows. Images represent ~40 cells that are in the mid-to-late division stages from two biological replicates. Adapted from Sharifzadeh, S., Boersma, M. J., Kocaoglu, O., Shokri, A., Brown, C. L., Shirley, J. D., et al. (2017). Novel electrophilic scaffold for imaging of essential penicillin-binding proteins in Streptococcus pneumoniae. ACS Chemical Biology, 12(11), 2849–2857. https://doi.org/10.1021/acschembio.7b00614.

PBP2b is the essential peripheral transpeptidase in S. pneumoniae (Tsui et al., 2016). Similar to PBP2x, it is a primary penicillin-resistance determinant (Grebe & Hakenbeck, 1996). We sought to assess the activity of PBP2b throughout the bacterial cell cycle using the lactone probe 8T, which labels PBP1b, PBP2x and PBP2b (Figs. 4, 7 and 8). In this regard, first we treated S. pneumoniae E193 cells with 7FL probe to saturate PBP2x, followed by 8T probe (Fig. 4). Using this method, we could simultaneously visualize the activity of PBP2x and PBP2b in live pneumococcal cells (Fig. 8).

Fig. 7.

8T labeling of S. pneumoniae cells with or without PBP1b and a PBP2x-inhibiting concentration of methicillin. Two cultures of WT (IU1945) and Δpbp1b (E193) were pretreated with 0.1μg/mL methicillin for 20min, followed by labeling with 8T. The top row contains images of different cells in each plane, while the bottom two rows contain pairs of images of the same cell, with rotations around the indicated axis with the white arrow. The absence of PBP1b does not affect the labeling of 8T, but pretreatment of methicillin results in elongated cells and a lack of septal labeling, as shown with the empty arrows. These images are representative of >40 cells at all stages of division for each condition from two biological replicates. Adapted from Sharifzadeh, S., Boersma, M. J., Kocaoglu, O., Shokri, A., Brown, C. L., Shirley, J. D., et al. (2017). Novel electrophilic scaffold for imaging of essential penicillin-binding proteins in Streptococcus pneumoniae. ACS Chemical Biology, 12(11), 2849–2857. https://doi.org/10.1021/acschembio.7b00614.

Fig. 8.

Labeling with 7FL followed by 8T reveals distinct PBP localization patterns for PBP2x and PBP2b. Wild-type (IU1945) and Δpbp1b (E193) S. pneumoniae were grown, labeled with 7FL, washed, labeled with 8T, and imaged. Each set is two rows of images of the same cell, with the bottom row rotated relative to the top row around the indicated axis. In both wild-type and PBP1b knock-out strains, 8T labeling is restricted to peripheral rings when the cells are labeled first with 7FL. Solid arrows highlight central septal labeling by 7FL, and empty arrows highlight the absence of 8T labeling from the center of the division site. For each condition, ~40 mid-to-late division cells from two biological replicates were imaged and representative results are shown. Reproduced from Sharifzadeh, S., Boersma, M. J., Kocaoglu, O., Shokri, A., Brown, C. L., Shirley, J. D., et al. (2017). Novel electrophilic scaffold for imaging of essential penicillin-binding proteins in Streptococcus pneumoniae. ACS Chemical Biology, 12(11), 2849–2857. doi:10.1021/acschembio.7b00614.

2.2.3. Live-cell imaging

Both WT and Δpbp1b cells labeled with (2R,3S)-β-lactone-L-Phe-fluorescein (7FL) showed that PBP2x localized at the division septum as a ring in early-to-mid divisional cells (Figs. 5 and 6). In mid-to-late divisional cells, PBP2x localized to both the constricting ring and a separate site at the center of the ring, as well as the equators of daughter cells. Finally, enzyme activity was constricted down to a single point at the division site before cell separation. This indicates that subpopulations of active PBP2x demonstrate different localization patterns during a single constriction event, which is in agreement with other reports of PBP2x localization (Land et al., 2013; Tsui et al., 2014), as well as newly synthesized PG labeled by FDAAs (Fig. 5). Pretreatment with the PBP2x-specific inhibitor methicillin (Land et al., 2013) revealed elongated cells with a labeled ring at the division site, but with minimal or no constriction evident (Fig. 6). Importantly, >98% of methicillin-treated late division cells in both the WT and Δpbp1b strains displayed empty septal rings, in contrast with untreated controls from both strains that showed central septal labeling in 85% of wild-type and 73% of Δpbp1b cells. This indicates a lack of active PBP2x due to methicillin inhibition, which aligns with previous results (Tsui et al., 2013) and strongly supports the conclusion that septal 7FL labeling is due primarily, if not exclusively, to PBP2x.

Fig. 5.

Labeling of PBP2x activity (7FL) co-localizes with regions of new cell wall synthesis as visualized with FDAAs. Wild-type (IU1945) and Δpbp1b (E193) S. pneumoniae cells were labeled with the hydroxycoumarin-amino-D-alanine (HADA, H, blue), followed by labeling with TAMRA-amino-D-alanine (TADA, T, red), and finally with 7FL (green). Each row contains six different views of the same cell. The white arrows indicate rotation of the cells around the indicated axis. New cell wall synthesis, as visualized by FDAA labeling, and PBP2x activity, as visualized by 7FL labeling, co-localize at all stages of division. For each condition, >40 cells at all stages of division from two biological replicates were imaged and a representative image is shown here. Adapted from Sharifzadeh, S., Boersma, M. J., Kocaoglu, O., Shokri, A., Brown, C. L., Shirley, J. D., et al. (2017). Novel electrophilic scaffold for imaging of essential penicillin-binding proteins in Streptococcus pneumoniae. ACS Chemical Biology, 12(11), 2849–2857. https://doi.org/10.1021/acschembio.7b00614.

PBP2b is the essential peripheral transpeptidase in S. pneumoniae (Tsui et al., 2016). In order to assess the activity of PBP2b throughout the cell cycle, WT and Δpbp1b cells were labeled with (2R,3S)-β-lactone-D-Phe-TAMRA (8T), which is specific for PBP1b, PBP2x and PBP2b (Fig. 4). Labeling with 8T yields a nearly identical labeling pattern as 7FL due to their shared targets (Fig. 7). Additionally, labeling with 8T after pretreatment with methicillin also mirrors the results seen with 7FL, reinforcing the idea that PBP2x alone is localized to the center of the constricting division site (Fig. 7).

Gel-based studies indicate that labeling Δpbp1b cells with 7FL, followed by labeling with 8T, individually tags PBP2x (green) and PBP2b (red) (Fig. 4). Imaging of dual labeled cells revealed that 7FL displayed a labeling pattern similar to the single labeling results (Fig. 8), as expected. However, labeling of 8T was excluded from the center of the division site in >83% of WT and Δpbp1b cells. Instead, it was restricted to the outer peripheral ring around the division site and the equatorial rings of future daughter cells, which have not yet begun to constrict. These results correlate well with previous work (Tsui et al., 2016) that demonstrated that PBP2b localization remained separate from and external to PBP2x during constriction. We have now shown that active PBP2b follows this pattern as well and ruled out the possibility that methicillin treatment disrupts PBP2b localization.

3. Gel-based analysis of PBP activity profile

3.1. Equipment

1. CO₂ incubator

2. Spectrophotometer (Genesys 20, Thermo Scientific or equivalent)

3. Plastic 17 × 100mm culture tubes (VWR, 60818–689)

4. Micropipettors

5. Micropipettor tips

6. 1.5-mL microcentrifuge tubes

7. Bench top microcentrifuge

8. Branson Sonifier 250 OR vial tweeter UP200St (Hielscher)

9. NanoDrop 1000 Spectrophotometer (Thermo Scientific) OR NanoPhotometer P330 (IMPLEN)

10. Analog heat block (VWR)

11. XCell SureLock mini-cell gel electrophoresis system (Thermo Fisher Scientific)

12. Typhoon 9210 gel scanner (Amersham Biosciences) equipped with the following filters: 510nm LP, 575 LP, BPB1 530DF20, and 665 LP

13. ImageJ software (NIH) for gel data analysis

3.2. Material

1. S. pneumoniae IU1945 (an unencapsulated derivative of serotype 2 S. pneumoniae strain D39), E177 (Δpbp1a) and E193 (Δpbp1b) cells (Lanie et al., 2007)

2. Brain heart infusion medium (powder, Bacto Brain Heart Infusion, VWR, 90003–038)

3. 1× Phosphate-buffered saline (PBS) (137mM NaCl, 2.7mM KCl, 10.1mM Na₂HPO₄, 1.8mM KH₂PO₄, pH 7.4; filter sterilized using a 0.22-μm filter)

4. 7FL and 8T probe (Sharifzadeh et al., 2017) stock solutions in DMSO (store at −80°C, protected from light; stable for years)

5. Bocillin-FL (1mg/mL, Invitrogen, cat. no. B-13233) in DMSO (store at −80°C, protected from light; stable for months)

6. Lysozyme (10mg/mL) in 1 PBS (from chicken egg white, Fluka)

7. 4× SDS-PAGE gel loading buffer (200mM Tris Cl, pH 6.8, 400mM dithiothreitol (DTT), 8% (w/v) SDS, 0.2% (w/v) bromophenol blue, 40% (v/v) glycerol)

8. BenchMark™ Fluorescent Protein Standard (Invitrogen, LC5928)

9. 10× Tris-glycine SDS running buffer (144.0g/L glycine, 30.0g/L Tris, 10.0g/L SDS; pH ~8.3)

10. Mini-polyacrylamide gels produced in-house [polyacrylamide gels were composed of 10% acrylamide resolving gel (10.5mL of 1.5M Tris-HCl buffer pH 8.8, 10.5mL of acrylamide:bis-acrylamide 29:1 (40% solution), 21mL of H₂O, 140μL of aqueous 10% ammonium persulfate (APS), 15μL of tetramethylethylenediamine (TEMED)) and 4.5% acrylamide stacking gel (2.5mL of 0.5M Tris-HCl buffer pH 6.8, 1.13mL of acrylamide:bis-acrylamide 29:1 (40% solution), 6.38mL of H₂O, 30μL of aqueous 10% APS, 10μL of TEMED)]

11. Coomassie brilliant blue R-250 staining solution (Bio-Rad, 1610436)

12. Additional reagents and equipment for SDS-PAGE

3.3. Protocols

3.3.1. Bacterial culturing

1. Frozen glycerol stocks of S. pneumoniae cells were inoculated into 5mL BHI broth in a plastic culture tube using one loopful with a sterile inoculating loop.

2. Serial dilutions were made by mixing each tube briefly and transferring 500μL to a tube containing 5mL fresh BHI broth. Six dilutions were made and tubes were incubated at 37°C in an atmosphere containing 5% CO₂ for 12–16h.

3. Following incubation, the optical density at 620nm (OD₆₂₀) of each tube was determined using a spectrophotometer.

4. A culture in exponential phase (with an OD₆₂₀ between 0.05 and 0.4) was diluted to an OD₆₂₀ of 0.02 in 2mL of fresh BHI broth prewarmed to 37°C and incubated statically without shaking at 37°C in an atmosphere of 5% CO₂.

5. When the culture reached an OD₆₂₀ of 0.2, cells were harvested by centrifugation.

3.3.2. β-Lactone probe labeling

1. Cell pellets from 1.5mL of S. pneumoniae cultures at OD₆₂₀ ~0.2 were harvested by centrifugation at 16,100 × g for 2min at room temperature (RT).

2. Pellets were resuspended with 1.0mL 1× PBS and centrifuged at 16,100 × g for 2min at RT.

3. Cell pellets were resuspended in 50μL of 1× PBS containing 1–30μg/mL of β-lactone probes and incubated in the dark at RT for 20min.

   Note: Incubation time could be optimized to enhance labeling efficiency, as needed.

   To prepare a reference sample, one washed pellet sample was resuspended in 50μL of 1× PBS containing 5μg/mL Boc-FL and incubated in the dark for 10min at RT.

4. The cells were pelleted using the above-described centrifugation settings.

5. Pellets were washed with 1.0mL 1× PBS.

6. Membrane fractions were collected for SDS-PAGE analysis (Section 3.3.4).

3.3.3. Dual labeling with 7FL and 8T

1. S. pneumoniae E193 (Δpbp1b) cells were grown in BHI broth at 37 °C in an atmosphere of 5% CO₂ to reach an OD₆₂₀ of 0.2–0.25, as described in Section 3.3.1.

2. Cell pellets from 1.5mL of culture were harvested by centrifugation (16,100 × g for 2min at RT).

3. Cell pellets were washed with 1mL of 1× PBS.

4. Cell pellets were resuspended in 50μL of 1× PBS containing 5μg/mL 7FL and incubated in the dark at RT for 20 min.

5. Cells were washed with 1mL 1× PBS.

6. Cells were suspended in 50μL of 5 μg/mL solution of 8T in 1× PBS and incubated in the dark for 20min at RT.

7. Cells were washed with 1mL 1× PBS.

8. Membrane fractions were collected for SDS-PAGE analysis (Section 3.3.4).

3.3.4. Collection of membrane fraction for SDS-PAGE analysis

1. Labeled cells from Sections 3.3.2 or 3.3.3 were resuspended in 100μL of 1× PBS containing 10mg/mL of lysozyme and were incubated for 30min at 37 °C.

2. The cells were lysed by a Branson Sonifier 250 (power setting 3, 30% duty cycle for 3 × 10s intervals) or vial tweeter UP200St (70% C, 95% A, 5% adjustment snap and 1s SD Interval/10s for a total of six 1min intervals with 1min cooling time in between, 12min total time) on ice.

3. Membrane proteome was isolated by centrifugation at 21,000 × g for 15min at 4°C.

4. Membrane proteome was resuspended in 100μL 1× PBS and the samples were homogenized by sonication (Branson Sonifier; power setting 1, 10% duty cycle for 1s, more cycles can be utilized if samples still appears cloudy. Alternatively, samples could be sonicated in a water bath sonifier).

5. Protein concentration was measured by NanoDrop 1000 spectrophotometer (Thermo Scientific) or NanoPhotometer P330 (IMPLEN). The protein concentration was adjusted to 5.0mg/mL by diluting with 1× PBS. For the blank solution, 1× PBS was used. (Note: An adjusted concentration within 2.5–5.0mg/mL would give a measurable in-gel fluorescent signal).

6. Thirty microliters of proteome sample were dispensed into a clean 1.5-mL microcentrifuge tube and 10μL of 4× SDS-PAGE loading buffer was added to each sample.

7. The samples were heated for 5min at 90–95°C to denature the proteins then cooled to RT.

8. Samples were analyzed by SDS-PAGE (Section 3.3.5).

3.3.5. SDS-PAGE analysis

Ten to twelve microliters of membrane sample from Section 3.3.4 were loaded onto a 10% acrylamide gel for SDS-PAGE (acrylamide: bis-acrylamide = 29:1). The protein bands were separated by gel electrophoresis for 1.5h, at 180V, 60W. The gel rig should be protected from light, to prevent photobleaching of the fluorophores. This was done by covering the gel rig with a cardboard box, but foil can also be used. The gel was rinsed with distilled water three times before fluorescence scanning (Section 3.3.6).

3.3.6. In-gel fluorescence detection of labeled proteins

After SDS-PAGE (Section 3.3.5), labeled proteins were directly visualized at 50-μm resolution in-gel using a Typhoon 9210 gel scanner with a 532nm laser and 580-nm bandpass filter for TAMRA and a 473nm laser and a 526-nm short-pass filter for fluorescein and BODIPY FL. All gel images were analyzed using ImageJ software (Schneider, Rasband, & Eliceiri, 2012). The background signal of the gel images was subtracted, and the brightness and contrast were manually adjusted to optimize the signal-to-noise ratio (all operations were performed over the entire gel uniformly). Following fluorescence gel analysis, gels were stained with Coomassie blue (using manufacturer’s protocol) and protein levels were assessed to account for differences in sample loading onto the gel.

4. In vivo imaging of PBP activity

4.1. Equipment

1. CO₂ incubator

2. Spectrophotometer (Genesys 20, Thermo Scientific or equivalent)

3. Bench top centrifuges (at RT and 4°C)

4. Bench top vortex mixer (VWR, VM-3000)

5. DeltaVision OMX 3D-SIM Super Resolution System (GE Healthcare)

6. DeltaVision OMX Software for image acquisition (GE Healthcare)

7. DeltaVision SoftWoRx Software for image processing (GE Healthcare)

4.2. Materials

1. S. pneumoniae cells. Cells are unencapsulated derivatives of D39, a serotype 2 strain. Wild-type (IU1945) (Lanie et al., 2007) and Δpbp1b (E193) (Land & Winkler, 2011).

2. Plastic 17 × 100mm culture tubes (VWR, 60818–689)

3. Glass 16 × 100mm culture tubes (Fisher, 1495925B)

4. Brain heart infusion (BHI) media (BD Bacto, 237500)

5. Microcentrifuge tubes

6. Dry ice

7. Round coverslips, 12mm diameter (Electron Microscopy Sciences, 72230–01)

8. Microscope slides (VWR, 16004–368)

9. 70% Ethanol

10. Lens paper (VWR, 52846–001)

11. Immersion oil, refractive index 1.518

12. 1 Phosphate-buffered saline (PBS) (Ambion, AM9625)

13. PBS containing 50mM glucose

14. GTE buffer (50mM glucose, 1mM EDTA, 20mM Tris-HCl, pH 7.5)

15. Vectashield Hardset Antifade mounting media (Vector Laboratories, 93952–26)

16. Methicillin (Sigma, 51454–50MG)

17. 7FL and 8T probe (Sharifzadeh et al., 2017) stock solutions in DMSO (stored at −80°C, protected from light; stable for years)

18. Fluorescent D-amino acids (hydroxycoumarin-amino-D-alanine, HADA and TAMRA-amino-D-alanine, TADA) were synthesized as reported (Kuru et al., 2012), with the following change: TADA was synthesized as reported for TDL, except that Boc-D-DAP-OH (N-alpha-t-butyloxycarbonyl-D-2,3-diaminopropionic acid) was used in place of Boc-D-Lys-OH (N-alpha-t-butyloxycarbonyl-D-lysine). 500mM stock solutions in DMSO were stored at −20°C and shielded from light

4.3. Protocols

4.3.1. Bacterial cell culture

Exponentially growing cultures of S. pneumoniae at OD₆₂₀ of 0.2 were prepared as described in Section 3.3.1.

4.3.2. β-Lactone probe labeling

1. Cells were pelleted as previously described and the supernatant discarded. Resuspend the cell pellet in 1mL of RT PBS with brief vortexing, and centrifuge for 5min at 16,000 × g at RT.

2. Discard supernatant and resuspend the cell pellet in either 50μL PBS with 5μg/mL 7FL or 100μL PBS with 5μg/mL 8T with brief vortexing.

3. Incubate the cells at RT for 20min in the dark.

4. After incubation, centrifuge for 5min at 16,000 × g at RT.

5. Discard supernatant and wash the cell pellet in either 1mL PBS (7FL) or 1mL PBS containing 50mM glucose (8T) with brief vortexing. Centrifuge for 5min at 16,000 × g at RT.

6. Discard supernatant and wash the cell pellet in either 200μL PBS (7FL) or 200μL PBS containing 50mM glucose (8T) with brief vortexing. Centrifuge for 5min at 16,000 × g at RT.

7. Discard supernatant and resuspend the cell pellet in 100μL cold GTE buffer with brief vortexing. Centrifuge for 5min at 16,000 × g at 4°C.

8. Discard supernatant and resuspend the cell pellet in 15μL Vectashield Hardset Antifade with brief vortexing. Keep cells on ice.

9. Clean coverslips and slides by applying 70% ethanol and wiping with lens paper.

10. Pipette 1.2μL of cells onto a coverslip and carefully place a microscope slide on top. Don’t move the slide, but put something over it (such as an empty tip box) to incubate in the dark for 15min at RT.

4.3.3. 3D-SIM (structured illumination microscopy) imaging

1. 3D-SIM imaging was performed using the DeltaVision OMX 3D-SIM Super Resolution System.

2. Image acquisition utilized DeltaVision OMX software.

3. The system was equipped with four Photometrics Cascade II EMCCD cameras that allow simultaneous imaging of up to four colors.

4. Laser lines used were 405 with emission filters of 419–465 (blue, for HADA), 488 with emission filters of 500–550 (green, for 7FL), and 561 with emission filters of 609–654 (red, for 8T or TADA). For all channels, exposure times were 5ms and %T was 100%.

5. For all samples, immersion oil with a refractive index of 1.518 was used. For each sample tested, a stack of 9–15 z sections (each 0.125μm thick) was rotated and visually checked for “halo” artifacts to ensure that the refractive index of the prepared sample matched the refractive index of the oil used.

6. Image processing utilized DeltaVision SoftWoRx software and consisted of OMX Reconstruction (Wiener filter value set at 0.001) and OMX Align.

4.3.4. Methicillin pretreatment

1. For pretreatment with methicillin followed by lactone probe labeling, make duplicate 2mL cultures of each strain in BHI broth (Section 4.3.1, step 4) and culture as described.

2. When the culture reaches an OD₆₂₀ of 0.12, add methicillin (final concentration of 0.1μg/mL) to one culture of each strain, and place the strains back in the incubator.

3. After 20min, add 1.5mL of the culture to a microfuge tube and centrifuge at RT for 5min at 16,000 × g.

4. Label and image cultures as described (starting with Section 4.3.2).

4.3.5. Dual β-lactone labeling

1. For dual labeling of cells with both lactone probes, the following changes were made after the 20min incubation with 7FL (Section 4.3.2, step 3).

2. After incubation, centrifuge for 5min at 16,000 × g at RT.

3. Discard supernatant and wash cell pellet in 1mL of PBS with brief vortexing. Centrifuge for 5min at 16,000 × g at RT.

4. Discard supernatant and resuspend cell pellet in 100μL PBS with 5μg/mL 8T with brief vortexing.

5. Incubate the cells at RT for 20min in the dark.

6. Wash and image as described above for β-lactone labeling (Section 4.3.2, starting at step 4).

4.3.6. FDAA β-lactone labeling

1. For labeling of cells with FDAAs and lactone probes, the following changes were made after making the 2mL culture in BHI broth (Section 4.3.1, step 4).

2. HADA stock solution (0.5 μL, stock 500mM, final concentration 125μM) was added to the 2mL culture and incubated as described (Section 4.3.1, step 4).

3. When the culture reached an OD₆₂₀ of 0.2, 0.5mL of the culture was added to a microfuge tube and centrifuged at RT for 5min at 16,000 × g.

4. Discard supernatant and resuspend cell pellet in 250μL prewarmed BHI broth containing TADA (final concentration 500μM) with brief vortexing.

5. Incubate cells at 37°C for 5min.

6. Cool cells on dry ice for exactly 20s, then centrifuge at 4°C for 2.5min at 16,000 × g.

7. Discard supernatant, resuspend cell pellet in 200μL ice-cold PBS with brief vortexing and centrifuge at 4°C for 2.5min at 16,000 × g.

8. Again, discard supernatant, resuspend cell pellet in 200μL ice-cold PBS with brief vortexing and centrifuge at 4°C for 2.5min at 16,000 × g.

9. Discard supernatant, resuspend cell pellet in 50μL RT PBS with 5μg/mL 7FL with brief vortexing.

10. Cells were labeled with 7FL, washed and imaged as described (Section 4.3.2, starting at step 3).

5. Additional considerations and controls

5.1. Probe concentration

Fluorescent probe concentrations were determined by measuring UV-Vis absorption of each solution at λ_max of its corresponding fluorophore. Probes were stored as DMSO solutions at −80 °C. For 7FL probe, 10× and 100× dilutions of the probe were made in 0.1M Tris-HCl (pH 8) and absorbance was read at 492nm using NanoPhotometer P330 (IMPLEN). The average of three absorbance values was used to calculate the concentration (ε = 78,000M⁻¹ cm⁻¹). The absorbance values must be read between 0.1 and 0.9 arbitrary units to ensure an accurate calculation. 8T probe was diluted in methanol, and concentration was calculated in the same way as for the 7FL probe (λ_max = 543nm, ε = 87,000M⁻¹ cm⁻¹).

5.2. Competition assays of penicillin V (Pen-V) and Boc-FL, 2F and 2T probes

In order to verify that the observed labeling is due to binding of the β-lactone probes to PBPs, we pretreated the cells with penicillin V, a well-known inhibitor of PBPs. To confirm the irreversibility of β-lactone labeling, labeled cells were treated with Pen-V after probe treatment.

In the Pen-V pretreatment assay, cells from 1.5mL of culture were harvested by centrifugation (16,100 × g for 2min at RT) and washed with 1mL of PBS, pH 7.4. Cell pellets were resuspended in 50μL of PBS containing 5μg/mL of Pen-V, and incubated for 30min at RT. Cells were pelleted and washed in 1mL of PBS. Next, the cells were resuspended in 50μL of PBS containing 5μg/mL Boc-FL, 5μg/mL 2F, or 5μg/mL 2T and incubated for 10min (Boc-FL) or 30min (2F, 2T) at RT. In the Pen-V post-treatment assay, cells were labeled with Boc-FL, 2F and 2T as described, washed and then incubated with 5μg/mL Pen-V. Finally, the cells were resuspended in 100μL PBS containing 10mg/mL lysozyme and incubated for 30min at 37°C. Cells were lysed and sample preparation for SDS-PAGE analysis was performed as described in Sections 3.3.4 and 3.3.5.

5.3. Methicillin pretreatment controls

Cells from 1.5mL of S. pneumoniae IU1945 and E193 cultures at exponential phase (OD₆₂₀ ~0.2) were harvested by centrifugation (16,100 × g for 2min at RT) and washed with 1mL of PBS, pH 7.4. Cell pellets were resuspended in 50μL of PBS containing 0.1μg/mL of methicillin and incubated for 30min at RT. Cells were pelleted and washed in 1mL of PBS. Next, the cells were resuspended in 50 μL of PBS containing 1 or 5μg/mL of 7FL or 5μg/mL Boc-FL (as control) and incubated for 20min at RT (10min at RT for Boc-FL). Cells were washed with 1mL of PBS and imaged, as described. In gel-based studies, the cells were resuspended in 100μL PBS containing 10mg/mL lysozyme and incubated for 30min at 37°C. Cells were lysed and sample preparation for SDS-PAGE analysis was performed as described in Sections 3.3.4 and 3.3.5.

6. Summary

Bacterial cell wall biosynthesis remains an outstanding target for new antibiotic development. However, there are significant gaps in our knowledge about this complicated process, including the mechanisms and control of PBP activity throughout cell growth and division. Continued development of tools to target and map the activities of the individual PBPs is a critical component to our understanding of bacterial growth and may lead to the identification of new therapeutic targets. Here, we report a methodology to profile the activity of individual PBPs in S. pneumoniae, using chemical probes based upon a scaffold not previously known to target the PBPs, the β-lactones. These probes demonstrated exquisite specificity for PBPs in bacterial cells and distinct interaction patterns with the PBPs in comparison to β-lactam antibiotics, which enabled us to examine the activity of two essential transpeptidases, PBP2x and PBP2b, in live S. pneumoniae. We found that PBP2b and PBP2x co-localize as a single ring in early division, but during mid-to-late division, PBP2x activity concentrates to the central septal site, while PBP2b activity remains at the outer division ring. This is the first report of visualization of the transpeptidation activity of individual PBPs in vivo, which confirmed that PBP2b and PBP2x display distinct patterns of localization during constriction. This methodology can be applied to any other bacterial system, as we have already tested in two model bacteria; B. subtilis and E. coli. As such, this work highlights the importance of the development of PBP-selective probes. Such probes, in combination with powerful cell biology tools, will illuminate the localization and activation of the PBPs in myriad bacteria.