Synthesis and Biological Evaluation of Biotinyl Hydrazone Derivatives of Muramyl Peptides

Authors

  • Didier Blanot,

    1. Univ Paris-Sud, Laboratoire des Enveloppes Bactériennes et Antibiotiques, Institut de Biochimie et Biophysique Moléculaire et Cellulaire, UMR 8619, Orsay, F-91405, France
    2. CNRS, Orsay F-91405, France
    Search for more papers by this author
  • Jooeun Lee,

    1. Department of Laboratory Medicine and Pathobiology, Medical Sciences Building, University of Toronto, Toronto, ON M5S 1A8, Canada
    Search for more papers by this author
  • Stephen E. Girardin

    Corresponding author
    1. Department of Laboratory Medicine and Pathobiology, Medical Sciences Building, University of Toronto, Toronto, ON M5S 1A8, Canada
    Search for more papers by this author

  • D.B. and J.L. contributed equally to this work.

Corresponding author: Stephen E. Girardin,stephen.girardin@utoronto.ca

Abstract

Muramyl peptides derived from bacterial peptidoglycan have long been known for their ability to trigger host innate immune responses, including inflammation and antimicrobial defense. Muramyl peptides have also been widely studied for their role as immune adjuvants. In mammals, the nucleotide-binding oligomerization domain (Nod) proteins Nod1 and Nod2 detect distinct muramyl peptide structures and mediate their biological activity. Because of the poor immunogenicity of these small peptidoglycan derivatives, research in this field is currently limited by the lack of reagents to track or immobilize specific muramyl peptides. We present here the generation and initial biological characterization of synthetic muramyl peptides covalently coupled to dansyl or biotinyl derivatives and demonstrate that biotinyl coupling on the muramyl moiety results in derivatives that can be tracked by immunofluorescence and maintain full biological activity, as observed by their capacity to trigger Nod signaling. Moreover, using digitonin-mediated permeabilization techniques on live cells, we also demonstrate that biotinylated muramyl peptides efficiently reach the host cytosol, where they activate Nod signaling. Therefore, these derivatives represent useful probes to study the cell biology and the biochemistry of host responses to muramyl peptides.

Abbreviations:
A2pm

diaminopimelic acid

Bio-Ahx

N-biotinyl-6-aminohexanoyl

Bio-Ahx-NHNH2

N-biotinyl-6-aminohexanoyl hydrazine

BMEM

Dulbecco’s modified Eagle’s medium

Dns

dansyl

Dns-NHNH2

dansyl hydrazine

FCS

fetal calf serum

MDP

muramyl dipeptide

MurNAc

N-acetylmuramoyl

Tri-A2pm

tripeptide l-Ala-γ-d-Glu-meso-A2pm

It is now well established that the innate immunity system acts through the detection of microbial motifs known as pathogen-associated molecular patterns (1). Among these, peptidoglycan and its hydrolysis products, muramyl peptides are sensed by the intracellular nucleotide-binding oligomerization domain (Nod) proteins Nod1 and Nod2. While Nod2 is a general sensor for all peptidoglycans because it recognizes the common muramyl dipeptide (MurNAc-l-Ala-d-Glu) motif (2–4), Nod1 senses the γ-d-Glu-meso-A2pm dipeptide motif found in the peptidoglycan of most Gram-negative bacteria, and the meso-A2pm amino acid must be in terminal position for Nod1-mediated detection (4–6). Following detection of specific muramyl peptides, Nod proteins trigger a wide array of inflammatory, antimicrobial and immune adjuvant responses (7). Using indirect approaches, we and others recently demonstrated that muramyl peptides are efficiently internalized by mammalian cells through clathrin-mediated endocytosis (8,9). However, the lack of suitably labeled muramyl peptides prevents direct analysis of their trafficking by immunofluorescence. In the present work, we have synthesized and used derivatives consisting of these compounds linked with a dansyl (Dns) or a biotinyl-6-aminohexanoyl (Bio-Ahx) probe through a hydrazone functionality. Dansylated muramyl peptides did not display strong-enough emission capacity for the use in immunofluorescence on live cells. By contrast, biotinylated muramyl peptides with the biotinyl group covalently linked to the muramyl group were endowed with interesting properties because (i) they demonstrated good biological activity for both Nod1 and Nod2 agonists in luciferase assays; (ii) in association with streptavidin-coupled fluorochromes, they proved to display sensitivity compatible with their use in immunofluorescence. Using digitonin-mediated permeabilization techniques on live cells, we also demonstrated that biotinylated muramyl peptides efficiently reach the host cytosol, where they activate Nod signaling. Together, biotinylated muramyl peptides appear to display an interesting versatility for their use in different cellular assays. These molecules will therefore prove to be useful for studying the cell biology and biochemistry of Nod signaling.

Methods and Materials

Compounds

Dansyl hydrazine (Dns-NHNH2), N-biotinyl-6-aminohexanoyl hydrazine (Bio-Ahx-NHNH2), and l-Ala-γ-d-Glu-meso-A2pm (Tri-A2pm) were purchased from Fluka (St. Louis, MO, USA), Pierce (Rockford, IL, USA), and InvivoGen (San Diego, CA, USA), respectively.

Muramyl peptides

MurNAc and MDP were purchased from Sigma (St. Louis, MO, USA) and InvivoGen, respectively. MurNAc-l-Ala and MurNAc-l-Ala-d-Glu were synthesized according to the published procedures (10,11). The other muramyl peptides were obtained from the corresponding UDP-MurNAc-peptides (12) by mild acid hydrolysis (0.1 m HCl, 100 °C, 10 min) (4). The residue resulting from the evaporation of the reaction mixture was used without purification for hydrazone formation.

Synthesis of dansyl hydrazones

To the muramyl peptide (100 nmol) dissolved in water (10 μL), 3% (w/v) trichloracetic acid (10 μL) and 1% (w/v) Dns-NHNH2 in acetonitrile (20 μL) were added. The mixture was stirred for 20 min at 65 °C and evaporated in vacuo. The product was purified by RP-HPLC.

Synthesis of biotinylated hydrazones

To dry muramyl peptide (150 nmol), 7.5 mm Bio-Ahx-NHNH2 in 30% acetonitrile (100 μL) was added. The mixture was evaporated to dryness, then dissolved in methanol/water/acetic acid 95:4:1 (50 μL; v/v). The mixture was stirred overnight at 60 °C and evaporated in vacuo. The product was purified by RP-HPLC.

Purification of hydrazone derivatives

The reaction mixture was taken up in 20% (v/v) methanol and injected onto a column (250 × 4.6 mm) of Econosphere C18 (5 μm; Alltech, Carquefou, France) equipped with a guard column of ODS-Hypersil (Thermo-Fisher Scientific, Illkirch-Cedex, France). Elution was performed at a flow rate of 0.6 mL/min with a gradient of methanol (eluent A: 20 mm ammonium acetate in water/methanol 4:1; eluent B, 20 mm ammonium acetate in water/methanol 1:4; gradient: 0% B from 0 to 10 min, 0–60% B from 10 to 50 min, 60–100% B from 50 to 55 min, 100% B from 55 to 60 min). The compounds were detected at either 247 nm (Dns derivatives) or 220 nm (Bio-Ahx derivatives). Peaks were collected manually, evaporated, taken up in water, and lyophilized. The compounds were finally dissolved in water/methanol 2:1 (225 μL; v/v) and stored at −20 °C.

Amino acid analysis

Samples were hydrolyzed in 6 m HCl at 95 °C for 16 h. After evaporation of the acid, the hydrolyzates were dissolved in 67 mm trisodium citrate-HCl (pH 2.2) and injected into a Hitachi L8800 analyzer equipped with a 2620MSC-PS column (ScienceTec, Courtaboeuf Cedex, France).

MALDI-TOF mass spectrometry

Positive spectra were recorded in the reflectron mode with delayed extraction on a Perseptive Voyager-DE STR instrument (Applied Biosystems, Carlsbad, CA, USA) equipped with a 337-nm laser. The compound (0.5 or 1 μL) was deposited on the plate, followed by 2,5-dihydroxybenzoic acid (1 μL at 10 mg/mL in 20 mm diammonium citrate). After evaporation of the solvents, spectra were recorded at an acceleration voltage of +20 kV and an extraction delay time of 200 nseconds. External calibration was performed using the calibration mixture of the Sequazime peptide mass standard kit (Applied Biosystems).

Cell culture and reagents

Human HEK293T and HeLa epithelial cell lines (American Type Culture Collection) were cultured in Dulbecco’s modified Eagle’s medium (DMEM), supplemented with 5% fetal calf serum (FCS) and 1% penicillin/streptomycin. Cells were maintained in 95% air, 5% CO2 at 37 °C. Endotoxin-free FCS and phosphate-buffer saline were from Wisent (St-Bruno, QC, Canada). FCS was used after heat inactivation at 56 °C for 30 min. All cell culture reagents and antibiotics were also from Wisent. Dynasore (3-hydroxy-naphthalene-2-carboxylic acid (3,4-dihydroxy-benzylidene)-hydrazide monohydrate) and digitonin were from Sigma.

NF-κB activation assays

Transfections were carried out using polyethylenimine (PEI; Polysciences Inc., Warrington, PA, USA) in HEK293T according to the manufacturer’s instructions. Briefly, cells were transfected overnight with 75 ng of NF-κB luciferase reporter plasmid (Igκ-luc; Invitrogen, Carlsbad, CA). The empty vector (pcDNA3; Invitrogen) was used to balance the transfected DNA concentration. The expression vector for human Nod2 (0.2 ng/well) was a kind gift from Dr Nuñez (University of Michigan, Ann Arbor, MI, USA). Following transfection, biotinylated ligands were added the next day for 6 h (unless specified) before performing luciferase measurements. For NF-κB activation assays in digitonin-permeabilized cells, HEK293T were incubated for 10 min at 37 °C with Nod1 ligands in isotonic digitonin buffer (6) (500 μL) with or without 10 μg/mL digitonin, and then placed in DMEM for 6 h. The dose of 50 nm was used for modified muramyl peptides because our previous results (13) have demonstrated that this dose typically gives a non-saturated and close to maximal (usually from 50% to 100%) activation of Nod1 or Nod2. For positive controls (MDP and Tri-A2pm), the dose of 10 μg/mL (approximately 20 μm) was used, as it typically provides maximal response in our assays.

Immunofluorescence

HeLa cells grown on glass coverslips were stimulated for 30 min with biotinylated muramyl peptide or biotinylated transferrin, as indicated. Next, 2 μg/mL Streptavidin-Alexa488 (Invitrogen) was added for 30 min. In some conditions, 1 μg/mL of directly coupled Transferrin-Alexa568 (Invitrogen) was added as a positive control, as indicated. Finally, cells were washed, fixed (4% paraformaldehyde, 15 min), and stained with DAPI to visualize nuclei. Images were taken using Zeiss Z-1 epifluorescence microscope with a 63× oil fluorescence objective and deconvolved using Volocity software (Quorum Technologies, Guelph, ON, Canada).

Results and Discussion

Muramyl peptides, obtained either by chemical synthesis or by mild acid hydrolysis of the corresponding UDP-MurNAc-peptides, were reacted either with Dns-NHNH2 or Bio-Ahx-NHNH2 (Figure 1). For Dns hydrazones 1, 3, and 5, a procedure adapted from those of Mopper and Johnson (14) and Hull and Turco (15) was used. For Bio-Ahx hydrazones 2, 4, and 6–10, the method of Leteux et al. (16) was followed. The compounds were purified using RP-HPLC. Theoretically, such hydrazones are mixtures of four isomers, two syn/anti isomers of the acyclic form and two α/β anomers of the cyclic form (17) (Figure 1). As a matter of fact, the desired compound appeared as a dissymmetrical peak or even as 2–3 overlapping peaks. The identity of the hydrazones was established using MALDI-TOF mass spectrometry: the positive-ion spectra displayed protonated or/and sodiated molecular ions (M + H)+ and (M + Na)+ consistent with the calculated molecular masses (Table 1). Yields, determined using quantitative amino acid analysis, were in general moderate; they were higher for the biotinyl derivatives (21–95%) than for the dansyl ones (10–21%) (Table 1).

Figure 1.

 Preparation of hydrazone derivatives 1–10. See Table 1 for the nature of R1.

Table 1.   Yields and mass spectrometry analysis of hydrazone derivatives
CompoundR1R2Yield (%)aMolecular mass
CalculatedbFound
(M + H)+(M + Na)+
  1. aCalculated from the amount of starting material (isolated muramyl derivative for compounds 1–6, or UDP-MurNAc-peptide for hydrazones 7–10) by quantitative amino acid analysis.

  2. bMonoisotopic molecular mass.

  3. cMain peak.

 1OHDns10540.19540.87562.89c
 2OHBio-Ahx57646.30647.13669.14c
 3l-AlaDns15611.23611.91633.95c
 4l-AlaBio-Ahx95717.34718.23740.26c
 5l-Ala-d-GluDns21740.27Not observed762.96
 6l-Ala-d-GluBio-Ahx45846.38847.73869.39c
 7l-Ala-γ-d-Glu-meso-A2pmBio-Ahx271018.461019.43c1041.43
 8l-Ala-γ-d-Glu-l-LysBio-Ahx35974.47975.50c997.74
 9l-Ala-γ-d-Glu-meso-A2pm-d-Ala-d-AlaBio-Ahx341160.541161.461183.47c
10l-Ala-γ-d-Glu-l-Lys-d-Ala-d-AlaBio-Ahx211116.551117.64c1139.64

We next tested the biological activity of these muramyl peptides and, to do so, first performed luciferase assays in HEK293T cells. In this cellular system, HEK293T cells were first transfected overnight with the Igκ-luci reporter plasmid, which encodes for the luciferase gene whose expression is driven by NF-κB elements on its promoter. Along with the Igκ-luci plasmid, cells were transfected with expression vectors encoding either Nod1 or Nod2, to potentiate the cellular responses through these Nod-like receptor proteins. Next, cells were stimulated for 6 h with 50 nm biotinylated muramyl peptides and lysed for luciferase analysis. We observed that only compound 7 (Rl-Ala-γ-d-Glu-meso-A2pm) was able to stimulate Nod1-dependent responses (Figure 2A), in agreement with the previously reported requirement of Nod1-activating muramyl peptides to contain a terminal meso-A2pm (4–6). Indeed, compound 9 (R1 = l-Ala-γ-d-Glu-meso-A2pm-d-Ala-d-Ala), which also has a meso-A2pm, but not in terminal position, was unable to trigger Nod1, as we previously observed (4). Compound 6 (R1 = l-Ala-d-Glu) was able to trigger Nod2-dependent responses (Figure 2B), which is in agreement with the known capacity of MurNAc-l-Ala-d-Glu (4) or MurNAc-l-Ala-d-Glu-NH2 (MDP) (2,3) to activate Nod2. Biotinylated compounds 4 (R= l-Ala), 8 (Rl-Ala-γ-d-Glu-l-Lys), and 10 (R1 = l-Ala-γ-d-Glu-l-Lys-d-Ala-d-Ala) were unable to stimulate either Nod1 or Nod2 (data not shown), in keeping with our previous results on unmodified muramyl peptides (4). Moreover, in dose–response experiments, we noted that compounds 7 and 6 were able to activate Nod1 and Nod2, respectively, with efficiencies that were similar to the ones of the non-modified muramyl peptides, thus showing that the addition of the biotinyl group to muramyl peptides does not significantly affect the stimulatory activity of these molecules (data not shown). Finally, we also noted that the dansylated compound 5 (Rl-Ala-d-Glu), but not the dansylated compounds 1 (R= OH) and 3 (Rl-Ala), was able to stimulate Nod2 activity in the luciferase assay (Figure 2C). Together, these assays demonstrate that our modified muramyl peptides are biologically active, that the additional moiety (Dns or Bio-Ahx) does not hamper their Nod-stimulating capacity and that they activate Nod proteins with the same peptidic sequence requirements as the natural muramyl peptides.

Figure 2.

 Biotinylated and dansylated muramyl peptides are biologically active. (A, B) HEK293T cells were transfected overnight with the reporter gene Igκ-luci to monitor NF-κB activation, together with either of the peptidoglycan sensors Nod1 (A) or Nod2 (B). The following day, cells were either left unstimulated or stimulated for 6 h with biotinylated muramyl peptides (compounds are the following: 2, Bio-Ahx-MurNAc; 9, Bio-Ahx-MurNAc-l-Ala-γ-d-Glu-meso-A2pm-d-Ala-d-Ala; 7, Bio-Ahx-MurNAc-l-Ala-γ-d-Glu-meso-A2pm; 6, Bio-Ahx-MurNAc-l-Ala-d-Glu) or with the positive controls Tri-A2pm [for Nod1, in (A)] or muramyl dipeptide (MDP) [for Nod2, in (B)]. (C) HEK293T cells were transfected overnight with Igκ-luci, together with Nod2. The following day, cells were either left unstimulated or stimulated for 6 h with dansylated muramyl peptides (compounds are the following: 1, Dns-MurNAc; 3, Dns-MurNAc-l-Ala; 5, Dns-MurNAc-l-Ala-d-Glu). All muramyl peptides used (biotinylated or dansylated) were added at the final concentration of 50 nm, and positive controls were at 10 μg/mL. NS, non-stimulated.

We next investigated whether dansylated or biotinylated muramyl peptides could be used for direct fluorescence or immunofluorescence, respectively, to follow their delivery and trafficking within host cells. HeLa cells grown on coverslips were first stimulated for 30 min with dansylated muramyl peptides. Because the dansyl group is intrinsically fluorescent (excitation λ = 336 nm; emission λ = 531 nm), direct visualization can be achieved using an epifluorescence microscope. Unfortunately, the fluorescence emitted by the muramyl peptides internalized into HeLa cells was under the detection limit (data not shown), probably due to the absence of an amplification step (typically provided by the sequential use of two antibodies).

We therefore repeated these studies using a biotinylated muramyl peptide, which has the advantage of allowing an amplification step through the use of fluorescent streptavidin-conjugated molecules interacting with the biotin group. HeLa cells were stimulated with compound 7 or Biot-transferrin as a control, followed by Streptavidin-Alexa488 together with Transferrin-Alexa568. Using this technique, we successfully detected Bio-Ahx-MurNAc-tripeptide inside HeLa cells and demonstrated that the molecule was colocalized with transferrin, a marker of early and recycling endosomes (Figure 3), in agreement with previous reports showing that muramyl peptides are internalized into mammalian cells through clathrin-mediated endocytosis (8,9).

Figure 3.

 Internalization of biotinylated muramyl peptides in HeLa cells. (A–C) HeLa cells grown on coverslips were first stimulated for 30 min with either medium (A), 1 μg/mL Bio-Ahx-MurNAc-l-Ala-γ-d-Glu-meso-A2pm 7 (B) or 1 μg/mL Biot-Transferrin (C). Next, 2 μg/mL Streptavidin-Alexa488 (Invitrogen) was added together with 1 μg/mL Transferrin-Alexa568 (Invitrogen) for 30 min. Finally, cells were washed, fixed (4% paraformaldehyde, 15 min), and stained with DAPI to visualize nuclei. Images were taken using Zeiss Z-1 epifluorescence microscope with a 63× oil fluorescence objective and deconvolved using Volocity software (Quorum Technologies). Transferrin-Alexa568 was used as a positive control in our experiments. Arrows indicate biotinylated muramyl peptide-containing vesicles. Tfn, transferrin. Biot, biotin. MP, compound 7. Strep, streptavidin.

We previously demonstrated that, following internalization into endosomes, Nod1-activating muramyl peptides were likely processed by host hydrolases in the lumen of early endosomes to generate muramyl-free peptides that can be transported to the cytosol through the oligopeptide transporter SLC15A4, and possibly other transporters (8). Therefore, when using a biotinylated muramyl peptide to stimulate host cells, the molecule would be cleaved in the early endosome to generate Bio-Ahx-MurNAc and free peptides, with the former remaining in the lumen of the endosome and only the latter reaching the cytosol. We reasoned that this could be a major limitation of the use of biotinylated muramyl peptides for projects aiming to study the fate of muramyl peptides in the host cytosol (such as the identification of muramyl peptide-interacting proteins by immunoprecipitation). Consequently, it was important to demonstrate that a procedure could be used, in which delivery of the biotinylated muramyl peptide to cytosolic Nod proteins would occur independently from endocytosis.

To circumvent this potential limitation to the use of muramyl peptides derivatives in live cells because of endosomal processing, we stimulated HEK293T cells with biotinylated muramyl peptides using a modified procedure that we have used previously for unmodified muramyl peptides (6), in which cytosolic delivery is direct and does not require endocytic trafficking. In this experimental set-up, HEK293T cells were first transfected overnight with the Igκ-luci reporter plasmid, along with either pcDNA3 or expression vectors encoding for Nod1 or Nod2. The following day, biotinylated muramyl peptides were added together with digitonin, a plant-derived toxin that destabilizes and permeabilize host membranes, thus allowing direct delivery of molecules to the cytosol. Accordingly, cells were pulsed for 10 min in a permeabilization medium containing digitonin plus biotinylated muramyl peptides, before replacing this buffer with regular cell culture medium for another 6 h prior to cell lysis and luciferase measurement. Of note, we first verified in our assays that digitonin-mediated delivery of the muramyl peptides was independent from endosomal trafficking, by showing that activation of Nod1-dependent signaling in digitonin-permeabilized cells by MurNAc-l-Ala-γ-d-Glu-meso-A2pm was insensitive to dynasore, a drug that efficiently inhibits clathrin-mediated endocytosis (data not shown).

Interestingly, we observed that these experimental conditions were also suitable for allowing delivery of biotinylated muramyl peptides to the cytosol, where compounds 7 and 6 could trigger the activation of Nod1 and Nod2, respectively (Figure 4). This demonstrates that the addition of the biotin-hydrazone arm to muramyl peptides does not affect the capacity of the molecules to access the cytosol through digitonin-mediated membrane permeabilization. Of note, in this experimental system, compound 7 was able to stimulate HEK293T cells endogenously (Figure 4A), which likely explains why this compound also displayed activity in Nod2-overexpressing cells (see Figure 4C). We did not observe such a capacity of biotinylated muramyl peptides to stimulate Nod pathways endogenously when trafficking through the endocytic machinery (see Figure 2), even though similar concentrations of muramyl peptides were used, which suggests that digitonin-mediated delivery allows for a more efficient internalization of muramyl peptides than natural endocytosis. Together, these data demonstrate that biotinylated muramyl peptides can be delivered to the host cytosol where they efficiently trigger Nod activation. This digitonin-based procedure will therefore be useful to study biochemical aspects of the activation of host cells by muramyl peptides, such as the nature of the protein complexes interacting with these bacterial molecules. Indeed, an important open question in the field of Nod biology is to determine whether Nod proteins directly interact with muramyl peptides or whether yet unknown adaptor molecules are required for detection and activation.

Figure 4.

 Enforced internalization of biotinylated muramyl peptides is sufficient to activate Nod1 and Nod2. (A–C) HEK293T cells were transfected overnight with the reporter gene Igκ-luci to monitor NF-κB activation, together with the empty vector pcDNA3 (A) or either of the peptidoglycan sensors Nod1 (B) or Nod2 (C). The following day, cells were transiently permeabilized using digitonin, in the presence or absence of biotinylated muramyl peptides (2, Bio-Ahx-MurNAc; 9, Bio-Ahx-MurNAc-l-Ala-γ-d-Glu-meso-A2pm-d-Ala-d-Ala; 7, Bio-Ahx-MurNAc-l-Ala-γ-d-Glu-meso-A2pm; 6, Bio-Ahx-MurNAc-l-Ala-d-Glu) or with the positive controls Tri-A2pm [for Nod1, in (A, B)] or muramyl dipeptide (for Nod2, in (C)) as indicated, and luciferase activity was measured 4 h post-stimulation (B–D). All biotinylated muramyl peptides used were added at the final concentration of 50 nm, and positive controls were at 10 μg/mL. NS, non-stimulated.

Conclusions and Future Directions

These experiments validate the approach of coupling muramyl peptides to the biotin probe. Using the biotin group as a bait, it can also be envisioned to use such derivatives to perform biochemical analyses of the proteins interacting with these bacterial molecules. Therefore, biotinylated muramyl peptides represent useful derivatives of Nod1/2 ligands, which will be critical for the understanding of how muramyl peptides activate the host innate immune system.

Acknowledgments

This work was supported by grants from the CNRS (UMR 8619), the ACI Microbiologie and the Biotox-Speednod project (to DB), the Crohn’s and Colitis Foundation of Canada as well as the Canadian Institutes for Health Research (to SEG).

Ancillary