DltE is annotated as a putative serine type D-Ala-D-Ala carboxypeptidase. However, sequence comparison with canonical members of this protein family revealed low sequence identity (Figure 1—figure supplement 1). Additionally, while the catalytic S-X-X-K motif (Motif 1) is conserved in DltE, the second (Y-X-S) and third (K/H-T/S-G) motifs that complete the active site are altered with only the catalytic Tyr and the Gly residues being conserved, respectively. This suggests a modified active site environment with potentially different catalytic and/or substrate binding mechanisms. Supporting this, we showed that the extracellular catalytic domain of DltE (DltE_extra 34–397) is not able to bind penicillin although this property is a hallmark of serine type D-alanyl-D-alanine carboxypeptidases that belongs to the PBP (penicillin binding proteins) protein family (Figure 1—figure supplement 2).

These features prompted us to determine the structure of the extracellular domain of DltE (DltE_extra) by X-ray crystallography (Supplementary file 1). The domain consists of two subdomains, an α-β sandwich (residues 76–126 and 246–397) that form a five strand (β1–β5) antiparallel β-sheet flanked by 5 α-helices (α2, α9–α12) and an α-helix-rich region (residues 127–245) folding in 6 α-helices (α3-α8) (Figure 1a and Figure 1—figure supplement 1). DltE is structurally homologous to the D-Ala-D-Ala carboxypeptidase R61 (DDCP) from Streptomyces (Figure 1b and c; Kelly et al., 1986; McDonough et al., 2002) that exhibits the classical β-lactamase fold of PBPs (Figure 1b). The positions of the active site S-X-X-K motif (¹²⁸SIQK¹³¹), containing the catalytic Ser128 and Lys131 residues, and the catalytic Tyr213 from the second motif are conserved (Figure 1a and b and Figure 2). These last two residues are expected to function as a general base during the acylation step and function as a relay mechanism for the transfer of a proton from the incoming substrate to the departing catalytic serine. In contrast, several striking features distinguish DltE from DDCP and more generally from PBPs and DD-carboxypeptidases. The β3-strand that usually bears the third motif and defines one side of the catalytic cavity is not conserved in terms of length and position (McDonough et al., 2002; Figure 1c). Second, DltE contains an extra α-helix (α1) at the N-terminus of the catalytic domain (Figure 1c), as well as three major structural differences affecting the regions lining the sides of the active site of DDCP and responsible for its cavity-like shape (Figure 1c, e and g): (1) a 15-amino acid long loop connecting α4 and α5 in DltE (residues 175–190 named Loop I on Figure 1c and d and Figure 1—figure supplement 1) that replaces a longer region of 20 residues that is folded in α-helices in DDCP; (2) the mostly unstructured segment ranging from residues 328–346 in DltE and located between α11 and β3 (Loop II on Figure 1c and d and Figure 1—figure supplement 1) is 10 residues longer in DDCP in which it forms an antiparallel β-sheet of three small strands that follows the main 5-stranded β-sheet; and (3) the loop between α8 and α9 (residues 257–283 in DltE) connecting the two subdomains is slightly shorter in DltE and adopts a different position (Figure 1c). This last difference is of particular interest because this loop was described as the Ω-like loop in β-lactamases and PBP proteins (Fetrow, 1995; Yi et al., 2016) and shown to play a key role in the maintenance of the active site topology and in the enzymatic activity (Figure 1e and Figure 1—figure supplement 1). The global architecture of the active site cavity of DltE is therefore reshaped and different from that of DDCP enzymes, adopting a deeper cleft topology largely opened on both sides of the protein (Figure 1d–g). Together with its inability to bind penicillin, such structural features strongly suggest that DltE possesses an alternative substrate recognition mode and/or a different substrate specificity than PBPs.

Figure 1 with 2 supplements see all

Download asset Open asset

Figure 1—source data 1

DltE 3D structure.

https://cdn.elifesciences.org/articles/84669/elife-84669-fig1-data1-v2.xlsx

Download elife-84669-fig1-data1-v2.xlsx

Figure 1—source data 2

Raw SDS-PAGE analysis of the purity of DltE_extra.

https://cdn.elifesciences.org/articles/84669/elife-84669-fig1-data2-v2.pdf

Download elife-84669-fig1-data2-v2.pdf

Figure 1—source data 3

Labeled SDS-PAGE analysis of the purity of DltE_extra.

https://cdn.elifesciences.org/articles/84669/elife-84669-fig1-data3-v2.pdf

Download elife-84669-fig1-data3-v2.pdf

Figure 1—source data 4

Size-exclusion chromatography (SEC) of the Ni-affinity purified DltE_extra on a Superdex 200 10/300 GL.

https://cdn.elifesciences.org/articles/84669/elife-84669-fig1-data4-v2.xlsx

Download elife-84669-fig1-data4-v2.xlsx

Figure 1—source data 5

Microscale thermophoresis (MST) normalized dose–response data for the binding interaction between penicillin and S. pneumoniae PBP2b.

The experiment was made in triplicate.

https://cdn.elifesciences.org/articles/84669/elife-84669-fig1-data5-v2.xlsx

Download elife-84669-fig1-data5-v2.xlsx

Figure 1—source data 6

Microscale thermophoresis (MST) normalized dose–response data for the binding interaction between penicillin and DltE_extra.

The experiment was made in triplicate.

https://cdn.elifesciences.org/articles/84669/elife-84669-fig1-data6-v2.xlsx

Download elife-84669-fig1-data6-v2.xlsx

Figure 2

Download asset Open asset

Figure 2—source data 1

DltE 3D structure in complex with tartare (TLA) or TCEP (Tris(2-carboxyethyl)phosphine hydrochloride).

https://cdn.elifesciences.org/articles/84669/elife-84669-fig2-data1-v2.xlsx

Download elife-84669-fig2-data1-v2.xlsx

Besides the apo form structure, we obtained two other DltE structures in complex with tartrate or TCEP (Tris(2-carboxyethyl)phosphine hydrochloride), two compounds arising from the crystallization conditions (Supplementary file 1 and Figure 2a and b). Intriguingly, these two molecules are anchored by an interaction network that involves eight different residues, among which are the conserved active site nucleophile Ser128 and the base Tyr213 (Ser 62 and Tyr 159 in DDCP; Figure 2a, b and e). Importantly, the catalytic Ser128 with the hydroxyl group of its side chain positioned at only 2.7 Å away from a carboxylic group of TCEP and tartrate is ideally placed as it would activate a substrate for the enzyme acylation event as described for DDCP (McDonough et al., 2002). This suggests that the tartrate and TCEP molecule may mimic the nature and position of the natural substrate of DltE in its active site and that the acylation/deacylation step of the catalytic reaction is probably preserved in DltE (Figure 2b and c). In addition, tartrate or TCEP established a series of interactions with the structural elements specific to DltE (see above) and not found in DDCP (Figure 2e). Notably, a Tyr residue (Tyr338 in DltE) interacting with TCEP or tartrate replaces the conserved arginine (Arg285 in DDCP) in carboxypeptidases. This arginine is key for the recognition of the terminal carboxylate of the peptide substrate and the carboxypeptidase activity (McDonough et al., 2002; Figure 2b, d and e). Furthermore, the interactions involving motif 2 and 3 that stabilizes the substrate in DDCP and in particular the penultimate D-Ala residue of the peptide substrate are also largely altered (Figure 2b, d and e). In DltE, the hydrophobic contact with TCEP or tartrate is only maintained with the conserved Tyr213 and Gly348 of motif 2 and motif 3, respectively. Instead, residues specific to DltE (Tyr338, His347, Arg345, and Leu349) stabilize the TCEP or tartrate molecules (Figure 2a and b). Taken collectively, our structural data suggest that DltE would not be able to recognize the D-Ala-D-Ala terminal end of a stem peptide as DDCP does and that DltE would not be a DD-carboxypeptidase.

In L. plantarum PG precursors, a D-Lac residue rather than the more common D-Ala is present in position 5 of the peptide stem with an ester bond between D-Ala⁴ and D-Lac⁵ (Ferain et al., 1996). Tartrate and TCEP molecules having chemical groups that are also found in D-Lac (Figure 3—figure supplement 1), we hypothesized that the DltE structural features described above could reflect its ability to act as a carboxylesterase trimming the terminal D-Lac in mature PG. However, two enzymes, namely, DacA1 and DacA2, homologous to the D-Ala-D-Ala-carboxypeptidase DacA (PBP3) of Streptococcus pneumoniae (Morlot et al., 2005) could be potentially responsible for the D-Lac hydrolysis. Thus, we generated an Lp^NC8 mutant deficient for dacA1 and dacA2 and analyzed the PG composition. As shown in Figure 3—figure supplement 2a, we detected the presence of disaccharide-depsipentapeptide (GlcNAc-MurNAc-L-Ala-D-Gln-mDAP-D-Ala-D-Lac). As D-Lac residues are not detected in WT strain (Bernard et al., 2011), this indicates that DacA1 and/or DacA2 could behave as carboxylesterases removing the terminal D-Lac. To confirm this, we purified the disaccharide-depsipentapeptide from the ∆dacA1∆dacA2 strain and measured the ability of the purified recombinant catalytic domains of DacA1 and DltE to catalyze D-Lac hydrolysis. We observed that DltE was not able to cleave the ester bond between D-Ala⁴ and D-Lac⁵ from the disaccharide-depsipentapeptide, whereas purified DacA1 efficiently did (Figure 3). These results indicate that DltE is not a carboxylesterase active on PG stem depsipentapeptide.

Figure 3 with 6 supplements see all

Download asset Open asset

Figure 3—source data 1

Microscale thermophoresis (MST) data for the binding of DltE_extra with purified lipoteichoic acid (LTA) from Lp^NC8.

https://cdn.elifesciences.org/articles/84669/elife-84669-fig3-data1-v2.xlsx

Download elife-84669-fig3-data1-v2.xlsx

Figure 3—source data 2

DltE has D-Ala esterase activity on lipoteichoic acid (LTA).

https://cdn.elifesciences.org/articles/84669/elife-84669-fig3-data2-v2.xlsx

Download elife-84669-fig3-data2-v2.xlsx

Since DltE is not active on PG and because dltE gene is part of the dlt operon in Lp^NC8, we hypothesized that DltE could have a D-Ala esterase activity on D-alanylated teichoic acids. To assess whether DltE could act on WTA and/or LTA, we first characterized their structure by NMR in Lp^NC8 and determined their alanylation levels. Multidimensional NMR spectroscopy analysis revealed that WTA from Lp^NC8 contained two main repetition units composed of ribitol (Rbo)-phosphate chains substituted by two α-Glc residues either in C-2 and C-4 positions (major unit) or in C-3 and C-4 positions (minor unit) (Figure 3—figure supplement 2b). Remarkably, no Ala substituents were detected. WTAs purified from ΔdltXABCD were structurally identical from those purified from Lp^NC8, which supports that the DltXABCD complex is not involved in the D-alanylation of WTA. In contrast to WTA, NMR analysis established that LTA from Lp^NC8 is constituted by repetitive units of glycerol (Gro)-phosphate chains that were either unsubstituted (Unit A) or substituted with either Ala (Unit B), α-Glc (Unit C) or Ala-6-αGlc (Unit D) groups at C-2 position of Gro residues (Figure 3b). Relative quantification of NMR signals associated with individual units showed that Unit A was the major form with A/B/C/D ratio of 62:19:15:4. NMR analysis of LTA extracted from ΔdltXABCD mutant confirmed the absence of Ala-substituted LTA, as demonstrated by the disappearance of GroB-2 and GroB1,3 signals, GlcD-6 signals and Ala-2 signals on the ¹H-¹³C HSQC spectrum and of GroB1 to 3 signals on the ¹H-¹H COSY spectrum (Figure 3—figure supplement 3). Relative quantification of signals demonstrated that LTA from ΔdltXABCD mutant consisted of a simple mixture of Unit A and C in a 70:30 ratio. Taken together, these results establish that Lp^NC8 cell envelope carries LTAs with a distinctive pattern of structure and substitution and that only LTAs are D-alanylated. In addition, these observations confirm that the Dlt machinery is necessary to the D-alanylation of Lp LTAs.

Given that only LTAs are alanylated in Lp^NC8, we started testing DltE ability to bind LTA through microscale thermophoresis (MST) experiments. As shown in Figure 3c, DltE_extra was able to efficiently bind LTA. Then, we examined the activity of DltE on D-Ala-LTA purified from Lp^NC8. To achieve this, we analyzed the relative amounts of Ala esterified to Gro (Gro-Ala) and free Ala released from LTA upon incubation with or without DltE_extra. Both compounds were detected in the control sample as a result of spontaneous release of D-Ala from LTA when the test was performed without DltE (Figure 3d and Figure 3—figure supplement 5). Upon incubation of LTA with DltE, we observed a clear decrease (56% decrease) of Gro-Ala and a concomitant increase of free D-Ala. This effect was enhanced (64% decrease) with a double amount of DltE_extra in the test. In contrast, no difference with the control sample was observed for LTA incubated with DltE^S128A catalytic mutant protein or with DacA1 (Figure 3d and Figure 3—figure supplement 4). These results establish that DltE possesses a D-Ala esterase activity that cleaves the ester bond between the substituting D-Ala and Gro in LTA chains.

To gain structural insights into this activity, we soaked DltE crystals with LTA extracted from the ΔdltXABCD mutant and determined the 3D structure of such enzyme-substrate complexes (Supplementary file 1). We observe that the cleft harbors an additional, elongated density compatible with the presence of a 2-mer unit of the LTA molecule (Figure 2c). Half of the electron density is located at the same place as the tartrate and TCEP ligands (Figure 2) and would correspond to one unit of Gro-phosphate. It is stabilized by the same interaction network described above for tartrate and TCEP that share similar chemical groups, that is, hydroxyl, carboxyl and phosphate with LTA (Figure 2d, Figure 3—figure supplement 1). The second half of the electron density extends further on the catalytic cleft of DltE in the vicinity of Glu290 (in helix α9) and Tyr351 (between β3 and β4), forming thus a negatively charged subsite. The latter is not conserved in DDCP and replaced by a hydrophobic subsite that was proposed to recognize the aliphatic portion of the peptide (Figure 3—figure supplement 6a and b). These structural features are in agreement with the ability of DltE to bind and hydrolyse D-Ala-LTA, and, together with the biochemical data, they confirm that DltE is a D-Ala esterase acting on D-Ala-LTA.

We previously reported that the machinery encoded by the dlt operon (including dltE) is necessary to support D-Ala esterification in Lp cell envelope and sustain Drosophila juvenile growth (Matos et al., 2017). As previously observed upon mild alkaline hydrolysis, D-Ala was released in appreciable amounts from Lp^NC8 (WT) cells¹⁴ (Figure 4a), whereas no D-Ala was released from an isogenic mutant deleted for the entire dlt operon (Matos et al., 2017). However, we observed a significant reduction (around 70%) of D-Ala esterified to LTAs in ΔdltE mutant cells and an even higher reduction (around 93%) for DltE^S128A catalytic mutant cells. These results indicate that the esterase activity of DltE, in addition to the activities of DltX, A, B, C, and D, contributes to the D-alanylation machinery of LTAs in bacterial cells.

Figure 4 with 1 supplement see all

Download asset Open asset

Figure 4—source data 1

The D-Ala esterase activity of DltE contributes to D-alanylation of the cell envelope and is required to sustain Drosophila juvenile growth.

https://cdn.elifesciences.org/articles/84669/elife-84669-fig4-data1-v2.xlsx

Download elife-84669-fig4-data1-v2.xlsx

Figure 4—source data 2

The D-Ala esterase activity of DltE is required to sustain Drosophila juvenile growth – replicates.

https://cdn.elifesciences.org/articles/84669/elife-84669-fig4-data2-v2.xlsx

Download elife-84669-fig4-data2-v2.xlsx

We next wondered whether an active DltE protein is necessary for L. plantarum support to Drosophila juvenile growth. To this end, we compared juvenile growth (larval size at day 6 after egg laying) and developmental timing (i.e. day of 50% population entry to metamorphosis) of germ-free animals and ex-germ-free animals associated at the end of embryogenesis with either WT or ΔdltE and dltE^S128A mutants. Both mutations did not affect the ability of Lp to thrive in the fly niche but significantly altered the ability of Lp to support larval growth and developmental timing (Figure 4b–d and Figure 4—figure supplement 1). These results demonstrate that DltE activity is required to sustain Drosophila juvenile growth and together with the structural and biochemical insights on DltE activity point to the importance of D-Ala-LTAs as cues supporting Drosophila growth.

In order to evaluate the importance of D-Ala-LTA for Drosophila juvenile growth, we undertook a genetic approach by generating L. plantarum strains either deprived of LTA through the deletion of ltaS encoding for the LTA synthetase (ΔltaS) (Gründling and Schneewind, 2007) or WTA through the deletion of tagO encoding for the enzyme catalyzing the first step of WTA biosynthesis (ΔtagO) (Andre et al., 2011). A similar amount of D-Ala esterified to the cell envelope was measured in ΔtagO mutant cells compared to WT cells, whereas a strong reduction of D-Ala (around 72%) was observed in ΔltaS mutant cells (Figure 5—figure supplement 1a). The absence of LTA chain synthesis in ΔltaS mutant was confirmed by western blot with anti-LTA monoclonal antibody (Figure 5—figure supplement 1b). These results confirm that LTA are D-alanylated but not WTA as determined above by NMR analysis of purified polymers (Figure 3). Of note, we hypothesize that the residual D-Ala released for ΔltaS cells might be attached to a glycolipid precursor of LTA as previously observed in Listeria monocytogenes (Webb et al., 2009).

We then tested the respective TA defective strains for their ability to support Drosophila juvenile growth (Figure 5a and Figure 5—figure supplement 2a). When associated with germ-free animals, WT and ΔtagO strains support optimal Drosophila juvenile growth while Δdlt_op or ΔltaS strains largely fail to do so despite colonizing well their host’s niche (Figure 5a, Figure 5—figure supplement 1c, and Figure 5—figure supplement 2a). These data demonstrate that LTA are important molecules mediating Lp’s support to Drosophila juvenile growth. Accordingly, presence of D-Ala esterified to LTA chains (in WT and ΔtagO strains) correlates well with strains ability to support Drosophila’s growth. Indeed, D-Ala-LTA presence is required for Lp-mediated Drosophila juvenile growth promotion phenotype (Figure 5a, Figure 5—figure supplement 1a and b, and Figure 5—figure supplement 2a). In the absence of D-Ala-LTA (in Δdlt_op) or LTA (in ΔltaS), Drosophila larvae are smaller (Figure 5a and Figure 5—figure supplement 2a). Taken collectively these data reinforce the notion that D-Ala-LTAs from Lp play a key role in supporting Drosophila juvenile growth.

Figure 5 with 2 supplements see all

Download asset Open asset

Figure 5—source data 1

D-Ala-LTAs are necessary bacterial cues supporting Drosophila intestinal response and juvenile growth.

https://cdn.elifesciences.org/articles/84669/elife-84669-fig5-data1-v2.xlsx

Download elife-84669-fig5-data1-v2.xlsx

Figure 5—source data 2

D-Ala quantification and persistence on fly niche of WT, ∆dlt_op, ∆ltaS, and ∆tagO strains.

https://cdn.elifesciences.org/articles/84669/elife-84669-fig5-data2-v2.xlsx

Download elife-84669-fig5-data2-v2.xlsx

Figure 5—source data 3

D-Ala-LTAs are necessary bacterial cues supporting Drosophila intestinal response and juvenile growth – replicates.

https://cdn.elifesciences.org/articles/84669/elife-84669-fig5-data3-v2.xlsx

Download elife-84669-fig5-data3-v2.xlsx

Figure 5—source data 4

Replicate western blot detection of lipoteichoic acid (LTA) in wild-type Lp^NC8 and mutant derivatives.

https://cdn.elifesciences.org/articles/84669/elife-84669-fig5-data4-v2.pdf

Download elife-84669-fig5-data4-v2.pdf

We had previously established that upon association with Lp, enterocytes from ex-germ-free Drosophila sense and signal the presence of Lp cells by at least two independent molecular mechanisms: (1) PGRP-LE-mediated mDAP-PG fragment recognition triggering Imd/Dredd signaling (Erkosar et al., 2015) and (2) sensing of bacterial cell envelope bearing D-alanylated teichoic acids and signaling by unknown host mechanisms (Figure 5b; Matos et al., 2017). Both signals were reported to be important for maximal intestinal peptidase expression and optimal support to growth promotion by Lp (Figure 5a, c and d and Figure 5—figure supplement 2a–c; Matos et al., 2017). To probe the respective contribution of LTA and WTA to this model, we tested the ability of ΔltaS and ΔtagO strains (respectively deprived of LTA or WTA) to support Drosophila growth and intestinal peptidase expression in a Drosophila genetic background (Dredd mutants flies) (Leulier et al., 2000) that blunts the host response to mDAP-PG fragments (Leulier et al., 2003), allowing us to focus on the contribution of the D-alanylation signal to Lp-mediated growth support and intestinal peptidase induction. As previously observed upon association with the Δdlt_op mutant, Dredd mutant larvae associated with ΔltaS are compromised in their juvenile growth potential (Figure 5a and Figure 5—figure supplement 2a) and their ability to stimulate intestinal peptidase expression (Jon66Cii and Jon65Ai; Figure 5c and d and Figure 5—figure supplement 2b and c). However, juvenile growth and intestinal peptidase expression were not markedly affected in ΔtagO-associated Dredd larvae (Figure 5a–d and Figure 5—figure supplement 2a–c). Importantly, we observed a cumulative effect of not having D-Ala-LTA (Δdlt_op) or not having LTA at all (ΔltaS) and altering host mDAP-PG sensing (compare WT -yw- and Dredd conditions, Figure 5a–d and Figure 5—figure supplement 2a–c). These results therefore support the notion that both signals are required for optimal support to larval growth by Lp and that in addition to mDAP-PG fragments, D-Ala-LTAs are likely an additional cue sensed by Drosophila enterocytes.

Next, we tested the hypothesis that D-Ala-LTAs act as direct signals sensed by Drosophila enterocytes. To this end, we purified the major components from the cell envelope of WT and ΔdltXABCD strains: WTA, LTA, and PG, and tested their ability to rescue Δdlt_op and ΔltaS-mediated larval phenotypes (Figure 6a–e and Figure 6—figure supplement 1a–d). First, germ-free, Δdlt_op or WT-associated animals were treated daily with the purified cell envelope components (WTA, PG, and LTA) for 5 d. On day 6, larvae were harvested and their size measured (Figure 6b–e and Figure 6—figure supplement 1a–d). Daily supplementation with purified WTA or PG either from WT or ΔdltXABCD strains did not impact the growth of GF or Δdlt_op-associated larvae (Figure 6b and c and Figure 6—figure supplement 1a and b). In contrast, the daily supplementation of purified LTA isolated from WT (D-Ala-LTAs) to larvae associated with the Δdlt_op strain shows a growth-promoting effect that leads to increased larval size in this condition compared to the non-supplemented control (Figure 6d and e). This growth-promoting effect was not observed in GF and WT-associated animals nor when Δdlt_op associated larvae were supplemented with non-alanylated-LTAs isolated from ΔdltXABCD strain (Figure 6d and e and Figure 6—figure supplement 1c and d). We repeated the similar experiment with ΔltaS-associated larvae and again observed an improved larval growth upon supplementation with LTA purified from WT (D-Ala-LTAs) but not from LTAs purified from ΔdltXABCD strain (i.e. non-alanylated-LTAs) (Figure 6e and Figure 6—figure supplement 1d). These results therefore establish that D-Ala-LTAs are necessary and sufficient to restore Drosophila juvenile growth.

Figure 6 with 2 supplements see all

Download asset Open asset

Figure 6—source data 1

D-Ala lipoteichoic acids (LTAs) are bacterial cues supporting Drosophila intestinal response and juvenile growth.

https://cdn.elifesciences.org/articles/84669/elife-84669-fig6-data1-v2.xlsx

Download elife-84669-fig6-data1-v2.xlsx

Figure 6—source data 2

Experimental replicates from Figure 6b–d.

https://cdn.elifesciences.org/articles/84669/elife-84669-fig6-data2-v2.xlsx

Download elife-84669-fig6-data2-v2.xlsx

Figure 6—source data 3

Experimental replicates from Figure 6f and g.

https://cdn.elifesciences.org/articles/84669/elife-84669-fig6-data3-v2.xlsx

Download elife-84669-fig6-data3-v2.xlsx

Finally, we repeated purified LTA supplementations on Dredd larvae associated with the ΔltaS strain for 5 d and dissected the midguts of size-matched larvae to analyze intestinal peptidase (Jon66cii and Jon65Ai) expression by RT-qPCR independently of any mDAP-PG sensing/Imd signaling input (Figure 6a, f and g and Figure 6—figure supplement 2). We observed that supplementation with D-Ala-LTAs was sufficient to recapitulate the effect of WT strain on intestinal peptidase expression even in the absence of any mDAP-PG signal (Figure 5b, Figure 6f and g, and Figure 6—figure supplement 2). Taken collectively, our results demonstrate that in addition to and independently of mDAP-PG signaling, D-Ala-LTAs are important direct bacterial cues supporting intestinal peptidase expression and Drosophila juvenile growth.

Regardless of the initial DltE in silico annotation as a PBP (PbpX2), our structural data revealed that the global architecture of its active site cavity is reshaped and different from that of canonical PBP enzymes suggesting an alternative substrate recognition mode and/or a different substrate specificity. Hence, we show that D-Ala LTA and not PG is the substrate of DltE. Given its atomic structure and D-Ala esterase activity, DltE may rather relate to Staphylococcus aureus FmtA, a D-amino esterase acting on teichoic acids (Rahman et al., 2016). FmtA is, however, different from DltE given that it shows residual PBP activity, is able to interact with ß-lactams, and is encoded apart from the dltABCD operon in S. aureus genome. So far, the hypothesis advanced for FmtA biological function concerns the removal of D-Ala from LTAs to make it available to transfer onto WTAs or removal of D-Ala from WTA to reset cell surface charge under particular conditions (Rahman et al., 2016). Considering that WTA are not alanylated in Lp^NC8, these observations suggest that DltE possesses a distinct function than FmtA. We notice an apparent discrepancy between DltE enzymatic activity and the D-alanine content of the dltE mutants (ΔdltE and dltE^S128A). We posit that DltE may be involved in the control of D-Ala distribution on the LTA chains and DltE inactivation would result in impaired functioning of the whole D-alanylation machinery. Our work paves the way to the characterization of the complete mechanistical framework in which the interplay between the Dlt machinery and DltE are required for the D-alanylation of LTAs.

We analyzed Lp^NC8 TAs composition through multidimensional NMR spectroscopy and determined their alanylation levels. Lp^NC8 LTAs are, as expected from this strain’s genomic information and previous Gram-positive bacteria LTA characterization (Percy and Gründling, 2014; Sutcliffe, 2011), constituted by poly-(Gro-phosphate) chains substituted, among others, with Ala. In contrast, but in agreement with the genetic information from Lp^NC8 genome, this strain produces poly(Rbo-phosphate) WTAs given the lack of the tag genes encoding the machinery responsible for the synthesis of poly-(Gro-phosphate) chains in its genome (Bron et al., 2012). The ability to produce poly-(Gro-phosphate) WTA chains is a strain-specific feature in Lp as the tag gene cluster is only found in 22 out of 54 strains recently studied at the genomic level (Martino et al., 2016). However, in the case of the Lp^WCFS1 strain, despite its genome carrying the genetic information to synthetize both types of WTA (poly-(Gro-phosphate) and poly-(Rbo-phosphate)), only poly(Gro-phosphate) WTAs were detected at steady state (Bron et al., 2012). Surprisingly, no Ala substituents were detected in Lp^NC8 WTAs. This is an original feature as in other L. plantarum strains such as Lp^WCFS1, WTA appears to be D-alanylated (Bron et al., 2012). Yet, these WTAs are composed of poly(Gro-phosphate) chains so it would be interesting to investigate the levels of D-alanylation in other Lp strains harboring only poly(Rbo-phosphate) WTAs. Taken together, these observations call for future systematic analysis of TA composition and modification among L. plantarum strains and question the potential correlation between WTA composition and/or TA D-alanylation patterns and the ability of the strain to benefit Drosophila growth.

We have previously established that PG fragments (Erkosar et al., 2015) and D-Ala esterification to the cell envelope (Matos et al., 2017) trigger intestinal peptidase expression and support Drosophila juvenile growth. Here, we report that in addition to PG fragments, D-Ala-LTA from Lp^NC8 cell envelop represent a bacterial cue necessary and sufficient for the ability of Lp^NC8 to promote Drosophila growth and upregulation of intestinal digestive peptidases in enterocytes. Given that L. plantarum is a gut luminal microbe and that the Drosophila midgut is lined with both thick chitinous matrix and thin mucus layer, yet permeable to large macromolecules and small nutrient particles (Lemaitre and Miguel-Aliaga, 2013), how exactly these macromolecules associated to the bacterial cell membrane reach the enterocytes remains elusive. Similarly to PG fragments, which are shaded in the extracellular media during bacterial cell division, LTA chains may also be shed after cleavage of the lipid anchor (Neuhaus and Baddiley, 2003). In addition, our recent findings report Lp^NC8’s ability to release microvesicles of multi-nanometers ranges (Grenier et al., 2021). This observation is coupled to previous reports showing that other Lactobacillus strains form microvesicles containing LTAs that are endocytosed by mammalian enterocytes (Champagne-Jorgensen et al., 2021), it is tempting to speculate that Lp^NC8 D-Ala-LTAs are released via bacterial microvesicles cargo. The Drosophila host would capitalize on these features by endocytosing cleaved LTA or such cargo and sense these bacterial cues through pattern recognition receptors expressed in enterocytes.

PG fragments are sensed intracellularly in the Drosophila midgut by PGRP-LE, which signals the Imd pathway to stimulate intestinal peptidase expression. How D-Ala-LTAs are sensed and signal in enterocytes remain elusive, but our results support the notion that another signaling cascade beyond the one engaged by PGRP-LE is leading to D-Ala-LTA-mediated intestinal peptidase induction. Taken together, it seems that Drosophila ensures an optimal intestinal response to bacterial cues by using additive rather than interdependent signals converging locally to intestinal peptidase expression and macroscopically to enhanced juvenile growth.

Given the strong analogies of the impact of L. plantarum strains on Drosophila and mouse growth (Schwarzer et al., 2016) and previous work focusing on LTA in host-Gram-positive bacteria interactions (Champagne-Jorgensen et al., 2021; Hara et al., 2018), an exciting perspective will be to test the contribution of D-Ala-LTA to L. plantarum-mediated host growth promotion in other symbiotic models including mammals.

Drosophila stocks were cultured as described in Erkosar et al., 2015. Briefly, flies were kept at 25°C with 12/12 hr dark/light cycles on a yeast/cornmeal medium containing 50 g/L of inactivated yeast. The low-yeast diet was obtained by reducing the amount of inactivated yeast to 7 g/L. Germ-free stocks were established as described in Erkosar et al., 2015. Axenicity was routinely tested by plating serial dilutions of animal lysates on nutrient agar plates. Drosophila y,w flies were used as the reference strain in this work. The following Drosophila line was also used: y,w,Dredd^F64 (Leulier et al., 2000).

Strains used in this study are listed in Supplementary file 4. Escherichia coli strains were grown at 37°C in LB medium with agitation. L. plantarum strains were grown in static conditions in MRS media at 37°C, unless differently stated. Erythromycin antibiotic was used at 5 μg/mL for L. plantarum and 150 μg/mL for E. coli in the context of the deletion strains construction.

Independent markerless deletions on ltaS, tagO, dacA1, and dacA2 genes of Lp^NC8 genome were constructed through homology-based recombination with double-crossing over as described by Matos et al., 2017 . Briefly, the 5′- and 3′-terminal regions of each region were PCR-amplified with Q5 High-Fidelity 2X Master Mix (NEB) from Lp^NC8 chromosomal DNA. Primers contained overlapping regions with pG+host9 (Maguin et al., 1996) to allow for Gibson Assembly. PCR amplifications were made using the following primers: XL01/XL02 and XL03/XL04 (ltaS), XL05/XL06 and XL07/XL08 (tagO), XL09/XL10 and XL11/XL12 (dacA1), XL13/XL14, and XL15/XL16 (dacA2) listed in Supplementary file 3. The resulting plasmids obtained by Gibson Assembly (NEB) were transformed into Lp^NC8 electrocompetent cells and selected at the permissive temperature (28°C) on MRS plates supplemented with 5 µg/mL of erythromycin. Overnight cultures grown under the same conditions were diluted and shifted to the non-permissive temperature (41°C) in the presence of 5 µg/mL of erythromycin to select single crossover integrants. Plasmid excision by a second recombination event was promoted by growing integrants at the permissive temperature without erythromycin. Deletions were confirmed by PCR followed by sequencing. The strain deleted for dacA1 and dacA2 was obtained by the sequential deletion of dacA1 followed by dacA2.

L. plantarum^NC8 strain carrying a modified version of the dltE gene (dltE^S128A) was built by knocking in the modified sequence on ΔdltE strain (Matos et al., 2017). dltE-modified sequence harboring the mutation S128A was obtained by PCR with modified primers (X19). The 5′- and 3′-terminal regions of dltE region together with the entire dltE gene were PCR-amplified with Q5 High-Fidelity 2X Master Mix (NEB) from L. plantarum^NC8 chromosomal DNA using primers XL17/XL18 and XL19/XL20 (Supplementary file 3). The two fragments were assembled with pG+host9 (Maguin et al., 1996). The resulting plasmid was transformed into ΔdltE electrocompetent cells and selected at the permissive temperature (28°C) on MRS plates supplemented with 5 µg/mL of erythromycin. Overnight cultures grown under the same conditions were diluted and shifted to the non-permissive temperature (41°C) in the presence of 5 µg/mL of erythromycin to select single crossover integrants. Plasmid excision by a second recombination event was promoted by growing integrants at the permissive temperature without erythromycin. dltE ^S128A knock-in was confirmed by PCR followed by sequencing.

Axenic adults were put overnight in breeding cages to lay eggs on sterile low-yeast diet. Fresh axenic embryos were collected the next morning and seeded by pools of 40 in tubes containing fly food. 1 × 10⁸ CFUs or PBS were then inoculated homogeneously on the substrate and the eggs. Fly tubes are incubated at 25°C until larvae collection. Drosophila larvae, 6 d after inoculation, were randomly collected and processed as described by Erkosar et al., 2015 . Individual larval longitudinal length was quantified using ImageJ software (Schneider et al., 2012). For the cell envelope components supplementation experiments, 1 µg of purified LTA, WTA, and PG (extracted from strains Lp^NC8 and ΔdltXABCD) resuspended in PBS were added, independently, to the fly food every day until day 5 (see Figure 6a). For comparison of components extracted from each of the two strains, the final LTA, WTA, and PG suspensions were adjusted to the same amount of Gro, Rbo, and Mur, respectively. At day 6, larvae were harvested and larval longitudinal length was quantified using ImageJ software (Schneider et al., 2012).

Axenic adults were placed in breeding cages overnight to lay eggs on sterile poor-yeast diet. Fresh axenic embryos were collected the next morning and seeded by pools of 40 in tubes containing fly food. A total of 1 × 10⁸ CFUs of each strain or PBS was then inoculated homogeneously on the substrate and the eggs and incubated at 25°C. The emergence of pupae was scored every day until all pupae had emerged. D50 (day when 50% of the pupae emerged) was determined using D50App (Matos et al., 2017).

To access bacterial CFU in the fly nutritional matrix, microtubes containing food and larvae were inoculated with 1 × 10⁷ CFUs of each strain, independently. Tubes were incubated at 25°C until being processed. For bacterial load quantification, 0.75–1 mm glass microbeads and 500 μL PBS were added directly into the microtubes. Samples were homogenized with the Precellys 24 tissue homogenizer (Bertin Technologies). Lysates dilutions (in PBS) were plated on MRS agar using the Easyspiral automatic plater (Intersciences). MRS agar plates were then incubated for 24 hr at 37°C. Bacterial concentration was deduced from CFU counts on MRS agar plates using the automatic colony counter Scan1200 (Intersciences) and its counting software. For larval loads, y,w axenic eggs were inoculated with 1 × 10⁸ CFUs of each strain and incubated at 25°C until collection. Size-matched larvae were harvested from the nutritive substrate and surface-sterilized with a 30 s bath in 70% ethanol under agitation and rinsed in sterile water. Pools of five larvae were deposited in 1.5 mL microtubes containing 0.75–1 mm glass microbeads and 500 μL of PBS.

Axenic y,w and y,w,Dredd embryos were inoculated with 1 × 10⁸ CFU of Lp^NC8, Δdlt_op, and ΔltaS strains independently or kept axenic. Larvae were size matched for the four conditions and harvested at mid-L3 larval stage. Alternatively, larvae were inoculated with each one of the strains mentioned above and supplemented daily with cell envelop purified components (LTA, WTA, PGN) from Lp^NC8 or Δdlt_op cells. RNA extraction of five replicates of six dissected guts for each condition was performed as described by Matos et al., 2017. RT-qPCR was performed using gene-specific primer sets (Supplementary file 3) as described by Matos et al., 2017. Results are presented as the value of ∆Ct^gene/∆Ct^rp49.

Data representation and analysis was performed using GraphPad PRISM 8 software (https://www.graphpad.com). A total of 3–5 biological replicates were used for all experiments performed in this study in order to ensure representativity and statistical significance. All samples were included in the analysis. Experiments were done without blinding. Two-sided Mann–Whitney’s test was applied to perform pairwise statistical analyses between conditions.

DNA fragments were amplified by polymerase chain reaction using L. plantarum cDNA as a template and oligonucleotides listed in Supplementary file 5. The DNA encoding the extracellular domain of wild-type DltE (DltE_extra) and the catalytic mutant DltE_extra-S128A were cloned into the NcoI and XhoI sites of the pET-28a(+) vector that expresses proteins fused to a C-terminal hexahistidine (His)6 tag (Supplementary file 6).

DltE_extra and DltE_extra-S128A were expressed in E. coli BL21 (DE3)-RIPL cells. Cells were grown in LB media at 37°C and induced with 0.5 mM isopropyl b-D-1-thiogalactopyranoside (IPTG) overnight at 18°C. The cells were then harvested by centrifugation and resuspended in lysis buffer (50 mM Tris, pH 7.5, 500 mM NaCl, 10% [v/v] glycerol, 1 mM dithiothreitol [DTT] 0.01 mg/ml lysozyme, 0.006 mg/mL Dnase/RNase, protease inhibitor). The resuspended cells were disrupted by sonication and centrifuged at 14,000 × g for 45 min. Proteins were purified by a first step of Ni-NTA affinity chromatography with the elution buffer (50 mM Tris pH 7.5, 300 mM NaCl, 1 mM DTT, 250 mM imidazole). The eluted fractions were then concentrated and buffer exchange was performed in a centrifugal filter unit with gel filtration buffer (50 mM Tris pH7.5, 100 mM NaCl, and 1 mM DTT). The proteins were finally applied to a Superdex 200 16/600 GL size-exclusion column (GE Healthcare) and eluted with gel filtration buffer.

Crystallization conditions were screened at 293 K using the sitting-drop vapor-diffusion method and commercial crystallization kits Crystal Screen 1 and 2, PEG/Ion 1 and 2 (Hampton Research). Further optimization screenings around conditions of initial hits were also performed. The crystallization drops (0.2 µL protein solution and 0.2 µL reservoir solution) were set up using a Mosquito crystallization robot and equilibrated against 70 µL reservoir solution. DltE_extra and DltE_extra-S128A were concentrated to around 10 mg/mL prior to crystallization. Diffraction quality crystals grew after around 1 wk in three different conditions: (1) 26% PEG MME 5K, 0.1 M Tris pH8.5, 0.15 M LiSO₄ and 30% ethylene glycol (named hereafter apo condition); (2) 26% PEG 3350, 0.1 M ammonium tartrate (tartrate condition); and (3) 20% PEG 3350, 0.2 M ammonium tartrate, and 0.1 M TCEP HCl (TCEP condition). Crystals obtained in condition 2 were also soaked for 1 d in the crystallization condition containing 10 mM of LTA (prepared as described below) (LTA condition).

Crystals grown in conditions 2 and 3 were cryoprotected prior to data collection by rapid soaking in mother liquor containing 20% (v/v) glycerol and for the TA-soaked crystals 10 mM of TA. Crystals were flash-cooled in liquid nitrogen and diffraction data were collected at cryogenic temperature (100 K) on beamlines ID23-2 and ID30B at the European Synchrotron Radiation Facility (ESRF, Grenoble, France) and on beamlines PROXIMA-1 and PROXIMA-2 at SOLEIL synchrotron (Gif sur Yvette, France).

Data were processed using the XDS package (Kabsch, 2010). The space groups, asymmetric unit contents, and diffraction resolutions obtained for each of the crystallization conditions are presented in Supplementary file 1. The structures were solved by molecular replacement using Phaser implemented in PHENIX (Adams et al., 2010). The PDB entry 3WWX (Arima et al., 2016) was used as starting model to solve the DltE_extra-apo structure derived from the crystals obtained in the crystallization condition 1. The other three structures (named DltE_extra-tartrate; DltE_extra-TCEP; and DltE_extra-LTA) were solved using the DltE_extra-apo as a starting model. Structures were refined using iterative rounds of COOT (Emsley et al., 2010) and PHENIX (Adams et al., 2010) or Refmac5 of CCP4 (Murshudov et al., 1997). The quality of the final structure was assessed with MOLPROBITY (Chen et al., 2010) before deposition at the Protein Data Bank under the accession codes 8AGR(DltE_extra-apo), 8AIK(DltE_extra-tartrate), 8AJI(DltE_extra-TCEP), and 8AKH (DltE_extra-LTA). Sequence alignments and structure images were generated with PyMOL (Schrödinger, LLC), UCSF ChimeraX (Pettersen et al., 2021), and ESPript and ENDscript (Robert and Gouet, 2014). Data collection and final refinement statistics are presented in Supplementary file 1.

Protein–ligand interactions were analyzed by MST (Jerabek-Willemsen et al., 2011). Buffer of purified and concentrated protein samples was exchanged on a desalting PD-10 column to labeling buffer containing HEPES 25 mM pH 7.5, NaCl 300 mM, Tween 20 0.05% (w/v). Proteins were then labeled with NHS red fluorescent dye according to the instructions of the RED-NHS Monolith NT Protein Labeling kit (NanoTemper Technologies GmbH, Munchen, Germany). After a short incubation of target-partner complex, the samples were loaded into MST premium glass capillaries and eight measurements were performed (four with target alone and four with target-partner complex) at 22°C. The assays were repeated three times for each binding check experiment. Data analyses were performed using Nanotemper Analysis software provided by the manufacturer.

D-Ala esterified to teichoic acids was detected and quantified as described by Kovács et al., 2006. Briefly, D-Ala was released from lyophilized whole heat-inactivated bacteria by mild alkaline hydrolysis with 0.1 N NaOH for 1 hr at 37°C. After neutralization, the extract was incubated with Marfey’s reagent (1-fluoro-2,4-dinitrophenyl-5-L-alanine amide; Sigma). This reagent reacts with the optical isomers of amino acids to form diastereomeric N-aryl derivatives, which can be separated by HPLC. Separation of the amino acid derivatives was performed on a C₁₈ reversed-phase column (Zorbax Eclipse Plus C18 RRHD 2.1 × 50 mm 1.8 µm Agilent) with an Agilent UHPLC 1290 system with a linear elution gradient of acetonitrile in 20 mM sodium acetate buffer pH 5.0. The eluted compounds were detected by UV absorbance at 340 nm. Quantification was achieved by comparison with D-alanine standards in the range of 50–2000 pmol. Mean values were obtained from three independent cultures with two injections for each.

LTA detection was done as previously described (Webb et al., 2009) with some modifications. Bacterial cells from a 2 mL overnight culture in MRS were added to lysing matrix tubes containing 0.1 mm silica beads (Lysing Matrix B, MP Biomedicals) and were broken using an MP Biomedicals FastPrep 24 Homogenizer (4.5 m/s intensity for 30 s), with samples being kept on ice. After centrifugation of the suspension at 200 × g for 1 min to remove glass beads, cell envelopes and cell debris were recovered by centrifugation at 20.000 × g for 15 min. The pellet was resuspended in sample buffer containing 2% SDS. The sample buffer volume was normalized on the culture OD600nm. Samples were heated at 95°C for 20 min and centrifuged at 20.000 × g for 5 min. Supernatants were loaded on a 15% SDS-polyacrylamide electrophoresis (SDS-PAGE) gel. LTA were detected by western blotting after transfer onto a PVDF membrane (Immobilon-P transfer membrane, Millipore) by incubation with an anti-LTA mouse monoclonal antibody specific for poly-glycerolphosphate chains (clone 55, Origene, AM26274LE-N) at 1/1000 dilution, followed by anti-mouse antibody coupled to horseradish peroxidase (Thermo Fisher Scientific) at 1/2000. Western blot was revealed by chemiluminescence detection using Pierce ECL Western blotting substrate (Thermo Fisher Scientific) and a ChemiDoc imaging system (Bio-Rad).

L. plantarum strains (500 mL) were grown overnight in MRS medium. Cell walls were prepared as described previously (Matos et al., 2017), with some modifications. pH of buffer solutions was kept ≤6.0 in the different steps to avoid potential D-Ala hydrolysis from teichoic acids. Briefly, bacteria inactivated by heat treatment were boiled in 5% SDS in 50 mM MES buffer pH5.5 for 25 min. After centrifugation for 10 min at 20.000 × g, the pellets were resuspended in 4% SDS in 50 mM MES buffer pH 5.5 and boiled again for 15 min. The pellet was washed six times with 10 mM MES pH 5.5 preheated at 60°C. An additional step was added consisting in shearing sacculi with glass beads. Pellets were resuspended in 1 mL of 10 mM MES pH 5.5, and the suspension was added to lysing matrix tubes containing 0.1 mm silica beads (MP Biomedicals). The cells were broken using an MP Biomedicals FastPrep 24 Homogenizer (4.5 m/s intensity for 30 s). After centrifugation of the suspension at 200 × g for 1 min to remove glass beads, insoluble material containing cell walls was recovered by centrifugation 20.000 × g for 15 min. The pellet was resuspended in 50 mM MES pH 6.0, further treated with pronase, trypsin, DNase, RNase, and lipase, and finally boiled in 4% SDS in 10 mM MES pH 5.5 for 15 min. The final pellet was washed four times with 10 mM MES pH 5.5 and twice with MilliQ H₂O to remove SDS traces. The purified cell walls (containing PG, WTA, and polysaccharides) was lyophilized and further used for WTA and PG purification.

Purified cell walls were treated with hydrofluoric acid (48%) for 19 hr at 0°C to remove WTA and polysaccharides. The remaining insoluble purified PG was washed twice times with 250 mM Tris-HCl pH 8.0 and four times with MilliQ H₂O and lyophilized.

WTA were extracted from purified cell walls with TCA as described previously (Tomita et al., 2009) with some modifications. Briefly, lyophilized cell walls (50 mg; prepared as described above, keeping pH < 6.0 to avoid D-Ala hydrolysis from WTA) were incubated with 1 mL of 10% TCA at 4°C for 48 hr under rotation. The suspension was then centrifuged at 20.000 × g for 20 min at 4°C and the supernatant was recovered. WTAs were precipitated by addition of 5 V of ethanol and incubation overnight at –20°C. The pellet was purified by resuspension in TCA 10% and precipitation with ethanol. The final pellet was washed twice with cold ethanol, and the pellet was resuspended in MilliQ H₂O and lyophilized. WTAs were resuspended in 100 mM ammonium acetate buffer pH 4.8 (buffer A) and purified by anion exchange chromatography on a 1 mL HiTrap Q HP column (Cytiva) equilibrated with buffer A. WTAs were eluted with a gradient from 0 to 100% buffer B (buffer A containing 1 M NaCl) in 20 min. Fractions were collected and the presence of WTA was assessed in microplates with thymol-H₂SO₄ reagent (Engelhardt and Ohs, 1987), allowing detection of hexoses substituents of WTAs. The WTA-containing fractions were pooled and dialyzed against 10 mM ammonium acetate pH 4.8 with Float-a-Lyzer G2 dialysis devices (cut-off 500–1000 Da) (Spectra/Por) and lyophilized.

L. plantarum strains (5 L) were grown overnight in MRS. Bacteria were harvested at 4.500 × g for 10 min. The pellet was resuspended in 5 mL of 20 mM ammonium acetate buffer pH 4.6 (around 20 OD/mL). Bacteria were broken with a Constant Systems Ltd Basic Z Cell Disrupter (CellD) at 2000 bars. Non-broken cells were discarded by centrifugation at 3.000 × g for 15 min. LTA was obtained by butanol extraction allowing to keep D-Ala esterified on LTAs as described previously (Gründling and Schneewind, 2007; Morath et al., 2001). The supernatant was recovered and butanol-1 (1:1 V/V) was added in a glass container. The mixture was stirred for 30 min at room temperature. Insoluble material was discarded after centrifugation at 13.000 × g for 20 min in PPCO tubes (Nalgene). The liquid phase was recovered and centrifuged again; the aqueous phase (lower one) containing LTA was recovered. The sample was treated with DNAse II (Sigma D8764) at 50 µg/mL for 2 hr at 37°C to degrade contaminating DNA and lyophilized. LTA was further purified by hydrophobic interaction chromatography (HIC) with a Hi-prep Octyl FF 16/10 column with an AKTA chromatography system. LTA was dissolved in buffer A (10 mM ammonium acetate pH 4.7 containing 15% propanol-1). Elution was performed with a gradient of buffer B (10 mM ammonium acetate pH 4.7 containing 60% propanol-1) from 0 to 100% in 1 hr. Fractions were collected and the presence of LTA was tested by dot blot. Aliquots were spotted on a PVDF membrane (Immobilon-P transfer membrane, Millipore) and incubated with anti-LTA monoclonal antibody (Origene, AM26274LE-N) at 1/1000 dilution, followed by anti-mouse antibody coupled to horseradish peroxidase (Thermo Fisher) at 1/2000. Dot blot was revealed by chemiluminescence detection using Pierce ECL Western blotting substrate (Thermo Scientific) and detection with a Chemidoc system (Bio-Rad). Fractions containing LTA were pooled, dialyzed against 10 mM ammonium acetate pH 4.6 with Float-a-Lyzer G2 dialysis devices (cut-off 500–1000 Da) (Spectra/Por 500–1000 Da cut-off), and lyophilized.

PG was quantified by analysis of muramic acid (Mur) content after acid hydrolysis as described previously (Matos et al., 2017). PG (400 µg) was hydrolyzed by 6 N HCl for 16 hr at 100°C under vacuum. Mur was quantified by high-performance anion exchange chromatography coupled with pulse amperometric detection (HPAEC-PAD) with an ICS5000 system (Thermo Scientific) and a Dionex CarboPac PA-1 anion exchange column (4 × 250 mm) (Thermo Scientific) with a guard column. Quantification was made with a standard curve of pure Mur (Sigma-Aldrich) (50–1000 pmol).

WTA composition was obtained after hydrolysis with hydrofluoric acid 48% at 0°C for 48 hr. After HF evaporation, samples were further treated with 4 M trifluoroacetic acid (TFA) for 3 hr at 110°C. Composition was determined by HPAEC-PAD (ICS5000 system, Thermo Fisher Scientific) with a CarboPac PA20 column (Dionex). Composition and amount were determined by comparison with standard amounts of alditols (glycerol, ribitol) and monosaccharides glucosamine, galactosamine, glucose, galactose, rhamnose, and ribose.

LTA composition was obtained after hydrolysis by hydrofluoric acid 48% at 0°C for 48 hr. After HF evaporation, samples were treated with 4 M TFA for 3 hr at 110°C. Composition was determined by HPLC analysis with an Aminex HPX-87C column (Bio-Rad) with a Waters 2414 refractive index (RI) detector using alditols (glycerol, ribitol), monosaccharides and phosphate standards. Quantification was performed by comparison with standard amounts of the same compounds. D-Ala esterified to LTA was detected after alkaline hydrolysis using Marfey’s reagent derivatization, as described above for whole cells.

For supplementation experiments in fly tests, 1 µg of each purified component (PG, WTA, and LTA) was used. For comparison of components extracted from Lp^NC8 and ΔdltXABCD, the final PG, WTA, and LTA suspensions were adjusted to the same amount of Mur, Rbo, and Gro, respectively.

LTA was purified after butanol extraction allowing to keep D-Ala esterified on LTAs as described above. WTA was purified after TCA extraction from cell walls as described above. pH was kept below 6.0 at all purification steps of both polymers to avoid hydrolysis of the ester bond between D-Ala and the teichoic acid backbone chains. Samples were solubilized in highly enriched deuterated water (99.96% deuterium; EurisoTop, St-Aubin, France) and lyophilized. This process was repeated twice. Data were recorded on a 9.4 T spectrometer and a 18.8 T spectrometer were ¹H resonated at 400.33 and 800.12 MHz, and ¹³C resonated at 100.2 and 200.3 MHz, respectively. All samples were inserted in 3 mm tubes with matching amounts of D₂O. Acetone was added as an internal standard, starting from a solution of 2.5  μL of acetone–10  mL of D₂O. All pulse sequences were taken from the Bruker library of pulse programs and then optimized for each sample. Spectral widths were 12 and 200 ppm for the proton and carbon observations, respectively. TOCSY was performed with various mixing times of 40–120 ms. Edited ¹H-¹³C HSQC spectra were recorded with 1536 data points for detection and 256 data points for indirect direction.

Disaccharide-depsipentapeptide (GlcNAc-MurNAc-L-Ala-D-Gln-mDAP-L-Ala-L-Lac) was purified from PG of dacA1dacA2 double mutant of L. plantarum^NC8. PG was digested with mutanolysin (Sigma-Aldrich), and the resulting soluble muropeptides were reduced by NaBH₄ as described previously (Bernard et al., 2011). The reduced muropeptides were separated by reverse-phase ultra-high-pressure liquid chromatography (RP-UHPLC) with a 1290 chromatography system (Agilent Technologies) and a Zorbax RRHD Eclipse Plus C18 column (100 by 2.1 mm; particle size, 1.8 µm; Agilent Technologies) at 50°C using ammonium phosphate buffer and methanol linear gradient. The eluted muropeptides were detected by UV absorbance at 202 nm. The peak corresponding to the disaccharide-depsipentapeptide was collected. Purified muropeptide was then incubated with 10 µg of purified DltE_extra or DacA1 in 50 mM Tris-HCl, 100 mM NaCl buffer overnight at 37°C. The reaction mixtures were analyzed by RP-UHPLC as described above and by LC-MS with an UHPLC instrument (Vanquish Flex, Thermo Scientific) connected to a Q Exactive Focus mass spectrometer (Thermo Scientific). Mass spectra were collected over the range m/z = 380–1400. Data were processed using Xcalibur QualBrowser v2.0 (Thermo Scientific).

Purified DltE_extra (100 and 200 µg), DltE^S128A_extra (100 µg), or DacA1 (100 µg) were incubated overnight at 37°C with 200 µg D-Ala-LTA purified from L. plantarum^NC8 in a final volume of 100 µL of 20 mM MES, 150 mM NaCl pH 6.0. As a control, 200 µg of D-Ala-LTA was incubated in the same conditions without enzyme. After lyophilization, samples were treated with 48% hydrofluoric acid for 48 hr at 0°C, allowing LTA depolymerization without release of D-Ala esterified on Gro residues. Gro-D-Ala subunits and free D-Ala released from LTA chains were analyzed by LC-MS using an UHPLC instrument (Vanquish Flex, Thermo Scientific) connected to a Q Exactive Focus mass spectrometer (Thermo Fisher Scientific). A reverse-phase column (Hypersyl Gold AQ C18 column; 200 by 2.1 mm; particle size 1.9 µm; Thermo Fisher Scientific) was used for separation. To enhance the retention and resolution of the column, heptafluorobutyric acid (HFBA) was used as an ion-pairing agent. Samples were diluted 20-fold in buffer A with 2  µL injected onto the column. Buffer A contained 0.2% HBFA in MilliQ H₂O and buffer B contained 0.2% HBFA in acetonitrile/MilliQ H₂O (80:20, V/V). Elution was performed at flow rate 0.3  mL/min using an isocratic step of buffer A for 3 min followed by a gradient to 5% of buffer B in 5 min. Mass analysis was performed in single ion monitoring (SIM) mode by following Ala and Ala-Gro with the respective m/z of 90.05 and 164.09. Ala and Gro-Ala were identified by their MS and MS² spectra.

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

The authors thank Octobre Clocher, Dali Ma, and Cathy Ramos for their contribution to the early stages of the bacteriology and Drosophila tasks of this project, Céline Freton for help in microscale thermophoresis, Laurane Bernelin for her contribution at the initial stages of the DltE production and crystallization, and Emmanuel Maes and Jessica L Davis for technical assistance. We acknowledge the SFR Biosciences (UAR3444/US8) for access to the Protein Science (crystallogenesis robots) and the ArthroTools platform (Drosophila facility) and support on the beamlines ID23-2 and ID30B at the European Synchrotron Radiation Facility (ESRF, Grenoble, France) and on PROXIMA-1 and PROXIMA-2A at SOLEIL synchrotron (Gif sur Yvette, France), the PAGes core facility (http://plateforme-pages.univ-lille1.fr) and UAR 2014 /US 41 (Plateformes Lilloises en Biologie & Santé) for providing the scientific and technical environment conducive to achieve this work. We acknowledge ChemSyBio (Micalis, INRAE, France) for access to mass spectrometry facilities. This work was funded by the collaborative grant ANR-18-CE15-0011 to FL, MPCC and CG. Work in the FL lab was also supported by a FRM grant (DEQ20180839196), a FINOVI starting grant to RCM and an ENSL Emergence grant to RCM. Work in the CG lab was also supported by the Fondation Bettencourt-Schueller.

© 2023, Nikolopoulos, Matos, Ravaud et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.