Bacterial Cell Wall Analogue Peptides Control the Oligomeric States and Activity of the Glycopeptide Antibiotic Eremomycin: Solution NMR and Antimicrobial Studies

pharmaceuticals

Article Bacterial Cell Wall Analogue Peptides Control the Oligomeric States and Activity of the Glycopeptide Antibiotic Eremomycin: Solution NMR and Antimicrobial Studies

László Izsépi 1,2,Réka Erdei 2, Anna N. Tevyashova 3, Natalia E. Grammatikova 3, Andrey E. Shchekotikhin 3 , Pál Herczegh 4 and Gyula Batta 2,*

1 Doctoral School of Chemistry, University of Debrecen, H-4032 Debrecen, Egyetem tér 1., Hungary; [email protected] 2 Department of Organic Chemistry, University of Debrecen, H-4032 Debrecen, Egyetem tér 1., Hungary; [email protected] 3 Gause Institute of New Antibiotics, 11 B. Pirogovskaya, 119021 Moscow, Russia; [email protected] (A.N.T.); [email protected] (N.E.G.); [email protected] (A.E.S.) 4 Department of Pharmaceutical Chemistry, University of Debrecen, H-4032 Debrecen, Egyetem tér 1., Hungary; [email protected] * Correspondence: [email protected]

Abstract: For some time, glycopeptide antibiotics have been considered the last line of defense against Methicillin-resistant Staphylococcus aureus (MRSA). However, vancomycin resistance of Gram- positive bacteria is an increasingly emerging worldwide health problem. The mode of action of glycopeptide antibiotics is essentially the binding of peptidoglycan cell-wall fragments terminating in the D-Ala-D-Ala sequence to the carboxylate anion binding pocket of the antibiotic. Dimerization of Citation: Izsépi, L.; Erdei, R.; these antibiotics in aqueous solution was shown to persist and even to enhance the antibacterial effect Tevyashova, A.N.; Grammatikova, N.E.; in a co-operative manner. Some works based on solid state (ss) Nuclear Magnetic Resonance (NMR) Shchekotikhin, A.E.; Herczegh, P.; studies questioned the presence of dimers under the conditions of ssNMR while in a few cases, higher- Batta, G. Bacterial Cell Wall Analogue Peptides Control the Oligomeric order oligomers associated with contiguous back-to-back and face-to-face dimers were observed in the States and Activity of the crystal phase. However, it is not proved if such oligomers persist in aqueous solutions. With the aid of 15 15 Glycopeptide Antibiotic Eremomycin: N-labelled eremomycin using N relaxation and diffusion NMR methods, we observed tetramers Solution NMR and Antimicrobial and octamers when the N-Ac-D-Ala-D-Ala dipeptide was added. To the contrary, the N-Ac-D-Ala or

Studies. Pharmaceuticals 2021, 14, 83. (N-Ac)2-L-Lys-D-Ala-D-Ala tripeptide did not induce higher-order oligomers. These observations are https://doi.org/10.3390/ph14020083 interesting examples of tailored supramolecular self-organization. New antimicrobial tests have also been carried out with these self-assemblies against MRSA and VRE (resistant) strains. Academic Editor: Amelia Pilar Rauter

Received: 1 December 2020 Keywords: glycopeptide; resistance; eremomycin; dimer; oligomer; N-Ac-D-Ala-D-Ala; ligand; Accepted: 18 January 2021 15N relaxation; diffusion; NMR Published: 22 January 2021

Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in 1. Introduction published maps and institutional affil- iations. If the fear that infections by both Methicillin-resistant Staphylococcus aureus (MRSA) and vancomycin-resistant enterococci (VRE) bacteria will spread and cause worldwide epidemics becomes true, we may return to the pre-penicillin age in which bacterial in- fection contributed significantly to morbidity rates. Vancomycin-resistant bacteria have adopted a simple, but powerful device for their survival: the essential D-Ala-D-Ala target Copyright: © 2021 by the authors. sequence normally found in vancomycin-sensitive bacteria has changed to D-Ala-D-Lac. Licensee MDPI, Basel, Switzerland. Experimental and theoretical studies have shown that vancomycin binds less strongly to This article is an open access article such targets [1]. The newer VanC type resistance arises from the sixfold-lower affinity distributed under the terms and D D D D conditions of the Creative Commons of vancomycin for acyl- -Ala- -Ser than for acyl- -Ala- -Ala [2]. There is an ongoing Attribution (CC BY) license (https:// interest in the modification of glycopeptide antibiotics [3] and search for new antibiotics [4]. creativecommons.org/licenses/by/ An in-depth review of the structural biology of molecular recognition by vancomycin an- 4.0/). tibiotics was published [5]. It is known that most vancomycin antibiotics exert an enhanced

Pharmaceuticals 2021, 14, 83. https://doi.org/10.3390/ph14020083 https://www.mdpi.com/journal/pharmaceuticals Pharmaceuticals 2021, 14, 83 2 of 14

cell-wall analogue ligand binding through the formation of back-to-back dimers [6–13] in aqueous solution. The non-covalent dimers persist in both the presence and absence of small peptide ligands. Bulk methods, such as microcalorimetry [14] and sedimentation equilibrium [15], should provide valuable details on the thermodynamics of the aggregates. However, using these methods, it is difficult to differentiate between dimers and possible higher-order oligomers. Thermal titration curves of the association between vancomycin and peptides suggested ligand- induced aggregation, inconsistent with a simple 1:1 stoichiometry [14]. X-ray studies of vancomycin [16] and balhimycin [17] indicate the presence of four copies of antibiotic molecules in an asymmetric unit of the crystal. These molecules form the known antiparallel back-to-back dimers [9]. However, a face-to-face dimer interface is also observed, suggesting the formation of a virtually infinite chain via inter-dimeric H-bonds. In the crystal phase, such a structure was observed with a small ligand, N-Ac-D-Ala [16]. Theoretically, larger ligands with different stoichiometry may also induce higher-order oligomers. It was suggested [5,16], that the antibiotic/peptide stoichiometry may be 1:1 or 2:1 depending on the type of the ligand. From X-ray studies, a putative model suggests that it is possible to detect ligand-mediated and dependent oligomerization in solution, when face-face dimers are formed, and a docking site for a third dimer is created that stabilizes the face-to-face interaction [18,19]. The crystal asymmetric unit contains six monomers [20] of vancomycin and similarly of balhimycin [21]. Other biophysical methods (SAXS, DLS) supported the presence of vancomycin- complexed N-Ac-D-Ala-D-Ala hexamers. In fact, surface plasmon resonance studies [22,23] generally reveal better correlation between the cell-wall analogue peptide binding constant and antimicrobial activity. Interestingly, inhibition of the activity of eremomycin with an access tripeptide (N-Ac)2-L-Lys-D-Ala-D-Ala required 10–50 times higher concentration than for vancomycin in competition experiments [24]. Remarkably, the potency of glycopeptide antibiotics against B. subtilis ATCC 6633 was tested in ”competition” experiments in the presence of increasing amounts of (N-Ac)2-L- Lys-D-Ala-D-Ala [25]. Eremomycin (Figure1) and dechloroeremomycin showed increasing activity in contrast to weakly dimerizing antibiotics such as vancomycin. This observation was explained with the ligand-enhanced dimerization mode (co-operativity) [13,26]. Using an 15N-labelled glycopeptide antibiotic for the first time [27], we proved the dynamic equivalence of the two sides of the dimeric eremomycin and the flexibility of the residue-3 Asn sidechain. We have also shown that the amides of the binding pocket exhibit the highest amide exchange rate, which could be a manifestation of the water strip effect presumably due to 2–3 amide bond rotation. 15N-labelling of glycopeptides provides an excellent tool for establishing the oligomeric state via the measurement of their global reorientational correlation time from 15N relaxation. Eremomycin forms a very strong homo-dimer among vancomycin antibiotics, and cooperativity between its dimerisation and ligand binding has been proved [25]. There is a strong correlation between the dimerisation constant (Kdim) and antibacterial activity (MIC) for many glycopeptide antibiotics [28]. Until now, glycopeptide antibiotic oligomers larger than dimers have never been detected in aqueous solution, except using the size-exclusion chromatography method in case of vancomycin [18]. In the present work, we explore the higher-order oligomeric states of eremomycin in aqueous solution using 15N-labelled and unlabelled forms bound to different cell wall analogue peptides. In addition, antimicrobial activities are also investigated. Pharmaceuticals 2021, 14, 83 3 of 14

Figure 1. TheThe structure of er eremomycin.emomycin. Labels Labels ee andand ff indicateindicate eremosamine; eremosamine; gg indicatesindicates glucopyranose glucopyranose sugar sugar units. units. Numbering accord to amino acid sequential order, xx standstand forfor αα-CHs,-CHs, yy standstand forfor carbonylcarbonyl carboncarbon atoms.atoms.

2.2. Results Results 2.1.2.1. NMR NMR Relaxation Relaxation and and Diffusion Diffusion Evidence Evidence of of Ligand-Dependent Ligand-Dependent Oligomerisation Oligomerisation of 1515N-LabelledN-Labelled Eremomycin Eremomycin InIn our our earlier earlier 15NN NMR NMR relaxation relaxation study study [27] [27] at at 280 280 K, we determined the global correlationcorrelation time time of of the the dimeric dimeric 1515N-eremomycinN-eremomycin in inacetate acetate buffer, buffer, which which was was found found to be to 3.7be 3.7ns. The ns. The1H-151NH- HSQC15N HSQC spectra spectra allowed allowed the separate the separate NH assignments NH assignments of the two of the halves two 6 ofhalves the ofasymmetric the asymmetric eremomycin eremomycin dimer dimer (Kdim (K dim= 3= × 3 ×10106/M)/M) [26]. [26 ].Furthermore, Furthermore, it it was was demonstrateddemonstrated that thethe additionaddition ofofN N-Ac--Ac-DD-Ala--Ala-DD-Ala-Ala could could slightly slightly enhance enhance the the activity activity of ofthe the antibiotic antibiotic in in vitro. vitro. 15 15 1 FigureFigure 22 showsshows thethe resultsresults of of 15NN relaxationrelaxation experimentsexperiments (T(T11,T, T22,, 15N-1HH NOE) NOE) [29] [29] followedfollowed by thethe Lipari-SzaboLipari-Szabo model-free model-free analysis analysis [30 [30,31],31] as as applied applied to to resolved resolved NH NH signals, sig- 2 nals,yielding yielding S order S2 order parameters parameters and reorientational and reorientational correlation correlation times in times the tetra- in the and tetra- octamers. and octamers.The global The correlation global correlation time is roughly time is roug doubledhly doubled in the octamers in the octamers (15.7 ns (15.7 vs. ns 6.9 vs. ns) 6.9 if ns)compared if compared to tetramers to tetramers (A similar (A similar factor offactor two isof foundtwo is between found between tetramers tetramers and dimers). and dimers).This is a reliableThis is a evidence reliable ofevidence the presence of the of presence two kinds of of two oligomers kinds of (tetra oligomers and octa) (tetra parallel and 15 octa)in the parallel solution. in Applyingthe solution. a ApplyingN-T2 relaxation a 15N-T ﬁlter,2 relaxation the octamer filter, signalsthe octamer canbe signals removed can 15 befrom removed the N-HSQC from the spectra 15N-HSQC (Figure spectra3) that (Figure helped 3) thethat signal helped assignments. the signal assignments. The spectra Theof the spectra oligomers of the overlayed oligomers with overlayed the dimer with (Figure the dimer4) clearly (Figure show 4) clearly twin signals show twin for each sig- nalsform for according each form to according dominant to back-to-back dominant back-to-back dimer constituents, dimer cons withtituents, asymmetry with dueasym- to metrydisaccharide due to conformationsdisaccharide conf [9].ormations The presence [9]. The of putativepresence ligandof putative mediated ligand face-to-face mediated face-to-facearrangements arrangements in oligomers in didoligomers not change did no thet change basic asymmetry,the basic asymmetry, as proven as by proven doubled by doubledNMR signals NMR throughout signals throughout all oligomeric all oligomeric spectra. spectra.

Pharmaceuticals 2021, 14, 83 4 of 14

1 3 5 Tetramer 6.9 ns 0.9 2 4 7 Octamer Pharmaceuticals 2021, 14, 83 6 15.7 ns 4 of 14 Pharmaceuticals 2021, 14, 83 4 of 14 0.8

0.7 1 3 5 Tetramer 6.9 ns 0.6 0.9 2 4 7 Octamer 6 15.7 ns 0.5 0.8

0.7 0.4 S2 order parameter order S2 0.6 0.3 0.5 0.2 0.4 0.1 parameter order S2 0.3 0 0.21 2 3 4 5 6 7 8 9 10 11 12 NH-sites in the oligomers 0.1 Figure 2. 15NH S2 order parameters from 15N relaxation for tetra and octameric, N-Ac-D-Ala-D-Ala 0 ligand-filled eremomycin1 2 3 (2784 5 K). 6In the7 bar8 plot9 the10 tetramer11 12 is pink, and the octamer is cyan coloured. Numbers on the top NH-sitesof bars in themean oligomers amino acid residues in the heptapeptide sequence. 15 2 15 Figure 2. 2. NH15NH S Sorder2 order parameters parameters from fromN relaxation 15N relaxation for tetra and for octameric, tetra andN -Ac-octameric,D-Ala-D-Ala N-Ac-D-Ala-D-Ala ligand-ﬁlledligand-filled eremomycin eremomycin (278 (278 K). In K). the In bar the plot bar the plot tetramer the tetramer is pink, andis pink, the octamer and the is octamer cyan is cyan coloured. Numbers Numbers on the on top the of top bars of mean bars amino mean acid amin residueso acid in residues the heptapeptide in the sequence.heptapeptide sequence.

15N−eremomycin tetramer loaded with NAcDADA T2−filtered experiment, tauc = 6.9 ns

ppm 15N−eremomycin tetramer loaded with NAcDADA w6 112 T2−filtered experiment,w4 tauc = 6.9 ns ppm 114 w6 w2112 w4 116 114 118 w2 116 120 118 w3 122 120 w5 w3 124 122

126 124 w5

126 128 w1? 130 128 w1? 130 132 w7 132 w7 134 134 109 8 7 ppm 109 8 7 ppm 15 Figure 3. N-T2 ﬁltered HSQC experiment displays exclusively the tetramer signals, since the faster spin–spin relaxation destroys the signals of the octamer. Labels w stand for amide NH groups;

numbers are according to sequential position. Signals are doubled because of the asymmetry of 15 FiguretheFigure 3. dimeric N-T 3. 15 building2N-T filtered2 filtered blocks. HSQC HSQC One experiment of experiment the w1 signals displays displays is too weak, exclusivexclusiv and thereforeelyely the the tetramer is tetramer labelled signals, with signals, a since since the faster the faster spin–spinquestionspin–spin relaxation mark. relaxation destroys destroys the the signals signals of of the the octamer.octamer. Labels Labels w w stand stand for foramide amide NH NHgroups; groups; numbersnumbers are according are according to sequentialto sequential position. position. SignalsSignals areare doubled doubled because because of the of theasymmetry asymmetry of the of the dimericdimeric building building blocks. blocks. One One of ofthe the w1 w1 signals signals isis too weak,weak, and and therefore therefore is labelled is labelled with with a question a question mark. mark.

Pharmaceuticals 2021, 14, 83 5 of 14

Superimposed spectra of dimeric (acetate) and oligomeric (DADA) 15N−labelled eremomycin Superimposed spectra of dimeric (acetate) and oligomeric (DADA) ppm 15N−labelled eremomycin

ppm d4 110 d4 110 d4* 112 o2* d2 d2* d4*d6* 114 112 o2* d6 d2 Asn d2* d6* 116 114 d6 Asn d1, (folded) 116 118 d1, (folded) t2*,t2 118 120 t2*,t2 o2 d3 120 o2 d3* d3 122 d3* 122 124 124 126 d5 126 d5* d5 128 d5* 128 130 130 132 132 134 134 d7*,d7 d7*,d7 136 136 10.5 10.010.5 9.510.0 9.09.5 8.59.0 8.08.5 8.07.5 7.57.0 7.06.5 ppm6.5 ppm

Figure 4. Overlayed 15N HSQC spectra of N-Ac-D-Ala-D-Ala ligand-ﬁlled oligomeric and dimeric 15 15 FigureFigure eremomycin.4. Overlayed 4. Overlayed LettersN HSQC N d, HSQC tspectra and spectra o standof N -Ac-of for ND-Ac- NH-Ala-D amide-Ala-D-AlaD groups-Ala ligand-filled ligand-filled in dimers, oligomeric oligomeric tetramers and and and dimeric dimeric octamers, eremomycin. Letters d, t and o stand for NH amide groups in dimers, tetramers and octamers, re- eremomycin.respectively. Letters Numbers d, t denote and o stand sequential for NH order. amid Asteriskse groups (*)in dimers, label amide tetramers NH groups and octamers, belonging re- to spectively.spectively. Numbers Numbers denote denote sequential sequential order. order. Asterisks Asterisks (*) label (*) label amide amide NH NH groups groups belonging belonging to to the the the same monomeric unit within an asymmetric dimer building block. Not all signals are labeled same monomericsame monomeric unit within unit within an asymmetric an asymmetric dimer dimer building building block. block. Not Not all allsignals signals are are labeled labeled bebe- because of signal overlap. cause ofcause signal of signal overlap. overlap.

2.2. Titration of Eremomycin with N-Ac-D-Ala-D-Ala and N-Ac-D-Ala 2.2. Titration2.2. Titration of Eremomycin of Eremomycin with withN-Ac- N-Ac-D-Ala-D-Ala-D-AlaD-Ala and and N-Ac- N-Ac-D-AlaD-Ala 13 13 D D TitrationTitration 13of of 1- CC NN-Ac--Ac-D-Ala--Ala-D-Ala-Ala [32] [32 ]into into 16 16 mM mM eremomycin eremomycin solution solution was was Titration of 1- C N-Ac-15D-Ala-D-Ala [32] into 16 mM eremomycin13 solution was monitored measuring15 N-HSQC spectra and by observing the13 C NMR spectra of monitoredmonitored measuring measuring 15N-HSQC N-HSQC spectra spectra and and by observingby observing the the 13C NMRC NMR spectra spectra of of the the ligand.the ligand. Saturation Saturation of eremomycin of eremomycin with the with dipeptide the dipeptide ligand was ligand proven was at proven the end at of theti- ligand.end Saturation of titration of eremomycin in accordance with with the published dipeptide bulk ligand binding was proven constants at the (2800/M) end of ti- [33], trationtration in accordance in accordance with withpublished published bulk bulk binding binding constants constants (2800/M) (2800/M) [33], [33], which which means means >which 95% occupancy means > 95% with occupancy the dipeptide with the ligand. dipeptide Using ligand. the DOSY Using NMR the DOSY titration NMR technique titration > 95% techniqueoccupancy (Figures with 5the and dipeptide6 ), we obtained ligand. a Using similar the ligand DOSY binding NMR afﬁnity titration (K =technique 1960/Mol). (Figures 5 and 6), we obtained a similar ligand binding affinity (Ke e= 1960/Mol). (FiguresN-Ac- 5 andD-Ala 6), showed we obtained moderate a binding similar to ligand eremomycin binding in chemicalaffinity shift(Ke = titration 1960/Mol). experi- N-Ac-D-Ala showed moderate binding to eremomycin in chemical shift titration experi- N-Ac-D-Ala showed moderate binding to eremomycin in chemical shift titration experi- mentsments (Figure (Figure 77,, KeKe == 2100/Mol),2100/Mol), similarsimilar toto thethe dipeptidedipeptide afﬁnity. affinity. Titration Titration of of eremomycin eremomy- ments with(FigureN-Ac- 7, DKe-Ala- = 2100/Mol),D-Ala was followedsimilar to by the15 Ndipeptide HSQC experiments affinity. Titration (Figure8 of). eremomycin with N-Ac-D-Ala-D-Ala was followed by 15N HSQC experiments (Figure 8). cin with N-Ac-D-Ala-D-Ala was followed by 15N HSQC experiments (Figure 8).

FigureFigure 5. 5. TitrationTitration of of NN-Ac--Ac-DD-Ala--Ala-DD-Ala-Ala into into eremomycin eremomycin as as monitored monitored by NMR diffusiondiffusion constantscon- stants(DOSY). (DOSY). Though Though the change the change of diffusion of diffusion coefﬁcient coefficient is negative is negati onve theon the y axis, y axis, they they are are shown shown as Figureas positive5. positive Titration values values of N for-Ac- for convenience convenienceD-Ala-D-Ala of of ﬁtting.into fitting. eremomycin as monitored by NMR diffusion constants (DOSY). Though the change of diffusion coefficient is negative on the y axis, they are shown as positive values for convenience of fitting.

Pharmaceuticals 2021, 14, 83 6 of 14

Pharmaceuticals 2021, 14, 83 6 of 14 Pharmaceuticals 2021, 14, 83 DIFFUSION NMR 6 of 14 ppm

EREMOMYCIN DOSY DIFFUSIONin 50 mM acetate NMR buffer 1.6 equivalent NAC−DADA added −8.0ppm EREMOMYCIN DOSY inDIFFUSION 50 mM acetate NMR buffer

ppm 1.6 equivalent NAC−DADA added −8.5−8.0 EREMOMYCIN DOSY in 50 mM acetate buffer

1.6 equivalent NAC−DADA added −8.0 −9.0−8.5

−8.5

−9.5−9.0

−9.0

−10.0−9.5

−9.5

−10.0 11 10 9 8 7 6 5 4 3 2 1 0 ppm −10.0

11 10 9 8 7 6 5 4 3 2 1 0 ppm 11 10 9 8 7 6 5 4 3 2 1 0 ppm Figure 6. Overlayed NMR diffusion (DOSY) experiments of eremomycin: red is in the absence of Figure 6. Overlayed NMR diffusion (DOSY) experiments of eremomycin: red is in the absence of ligand; ligand; blue is blue after is aftercoaddition coaddition of 1.6 of 1.6 equivalent equivalent NN-Ac--Ac-DD-Ala--Ala-DD-Ala.-Ala. On On the the log10-based log10-based vertical vertical scale, scale,Figure aFigurea 0.1 0.16. Overlayed unit 6. Overlayed differencedifference NMR indicatesNMR indicates diffusion diffusion a factor a (DOSY)factor (DOSY) of two of experiments two forexperiments the for molecular the molecular of eremomycin: mass. mass. red red is isin inthe the absence absence of of ligand;ligand; blue blueis after is after coaddition coaddition of of1.6 1.6 equivalent equivalent N N‐Ac-Ac-‐DD‐-Ala-Ala‐D-Ala.‐Ala. On On the the log10-based log10‐based vertical vertical scale,scale, a 0.1 aunit 0.1 15N unitdifference eremomycin difference indicates (16 indicates mM) titration a factora factor with of NAC-D-Alaof two two for for thethe molecular mass. mass. 6 shift / (limiting-no ligand) sum of six NH15N eremomycin15N eremomycin (16 (16mM) mM) titration titration with with NAC-D-Ala NAC-D-Ala 6 6 Binding constant = 2100 / Mol 5 shift / (limiting-noshift / (limiting-no ligand) ligand) pH = 4.5, phosphate buffer sum of six NH sum of six NH Binding constant = 2100 / Mol 5 Binding constant = 2100 / Mol 5 pH = 4.5, phosphate buffer 4 pH = 4.5, phosphate buffer

4 4 3

3 3 2 2 3 equivalents added 2 1 3 equivalents added 1 3 equivalents added 1 0 0 0.0050 0.01 0.015 0.02 0.025 0.03 0.035 0.04 0.045 0 0.005 0.01 Added0.015 Volume0.02 [mL]0.025 0.03 0.035 0.04 0.045 Added Volume [mL] 0 0 0.005 0.01 0.015 0.02 0.025 0.0315 0.035 0.04 0.045 Figure 7. Chemical shift titrationAdded Volume of dimeric [mL] N-labelled eremomycin with N-Ac-D-Ala. Figure 7. Chemical shift titration of dimeric 15N-labelled15 eremomycin with N-Ac-D-Ala. Figure 7. Chemical shift titration of dimeric N-labelled eremomycin with N-Ac-D-Ala. y y Figure140 7.140 Chemical shift titration of dimeric 15N‐labelled eremomycin with N‐Ac‐D‐Ala.

y Tetramer,Tetramer, all all 120140 120

Tetramer, all 100 100120

80 80100 Octamer, all 60 Octamer, all 6080

Volume integral in HSQC expt.integral Volume HSQC in Octamer, all 4060 Volume integral in HSQC expt.integral Volume HSQC in Tetramer, w2 20 Tetramer, w2 2040 Volume integral in HSQC expt.integral HSQC Volume in Octamer, w2 0 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 Added volume of 100 mM NAc-DADATetramer,Octamer, [ ml w2 ]w2 020 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 AddedD volumeD of 100 mM NAc-DADA [ ml ] Figure 8. NN-Ac--Ac- D-Ala--Ala-D-Ala-Ala titration titration into intoOctamer, eremomycin w2 solutionsoluti on at 278278K, K, followed by measuring the 0 0integral0.01 volumes0.02 of0.03 tetra 0.04 and octameric0.05 0.06 NH signals.signals.0.07 0.08 0.09 Figure 8. N-Ac-D-Ala-D-Ala titration into eremomycin solution at 278K, followed by measuring the Added volume of 100 mM NAc-DADA [ ml ] integral volumes of tetra and octameric NH signals. Figure 8. N‐Ac‐D‐Ala‐D‐Ala titration into eremomycin solution at 278K, followed by measuring the integral volumes of tetra and octameric NH signals.

Pharmaceuticals 2021, 14, 83 7 of 14 PharmaceuticalsPharmaceuticals2021 2021, 14, 14, 83, 83 77 of of 14 14

FigureFigure 8 shows8 shows that that in intoto, toto, ca. ca. two two times times more more tetramers tetramers are generated thanthan octamersoctamers Figure 8 shows that in toto, ca. two times more tetramers are generated than octamers duringduring NN-Ac--Ac-DD-Ala--Ala-DD-Ala-Ala titrationtitration ofof eremomycin.eremomycin. However,However, exclusiveexclusive observationobservation ofof w2w2 during N-Ac-D-Ala-D-Ala titration of eremomycin. However, exclusive observation of w2 signalssignals mightmight be misleading, suggesting suggesting that that octa octamersmers are are formed formed first. ﬁrst. This This effect effect may may be signals might be misleading, suggesting that octamers are formed first. This effect may be bedue due to towater water saturation saturation impacts impacts at atthe the w2 w2 site site,, which which is isthe the most most sensitive, sensitive, being being close close to due to water saturation impacts at the w2 site, which is the most sensitive, being close to tothe the binding binding site. site. Solvent Solvent accessibility accessibility of of the the possible possible binding binding sites sites can bebe trackedtracked fromfrom the binding site. Solvent accessibility of the possible binding sites can be tracked from waterwater saturationsaturation differencedifference experimentsexperiments (Figure(Figure9 )9) that that locate locate the the binding binding sites. sites. water saturation difference experiments (Figure 9) that locate the binding sites. y 0.9 y 0.9 2 2 0.8 2* bound - unbound state to NAc-D-Ala 0.8 2* bound - unbound state to NAc-D-Ala 0.7 0.7 0.6 0.6 0.5 0.5 3 3* 3 3* 0.4 0.4 0.3 0.3 4 0.2 4 0.2 4* 0.1 4* 5* 5 0.1 5* 6 7 Asp NH2 5 6* 7* 6* 6 7 Asp NH2 0 7* 0

−0.1 −0.1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 1 2 3 4 5 6 Residues7 8 9 10 11 12 13 14 Residues FigureFigure 9.9. DifferencesDifferences in in water water saturation saturation efficien efﬁciencycy between between the the free free eremomycin eremomycin dimer dimer and and N- Figure 9. Differences in water saturation efficiency between the free eremomycin dimer and N-Ac-D D-Ala loaded dimer. Bars close to 1 unit mean huge differences in water accessibility due to Ac-N-Ac--AlaD-Ala loaded loaded dimer. dimer. Bars Bars close close to to 1 1 unit unit mean mean huge huge differencesdifferences inin waterwater accessibility due due to to ligand binding, while values close to 0 mean no change in water access to that site upon binding. ligandligand binding,binding, whilewhile valuesvalues closeclose to 0 mean no no chan changege in in water water access access to to that that site site upon upon binding. binding. Asterisks indicate the same monomeric units within dimers. Numbers on the top of bars mean AsterisksAsterisks indicateindicate thethe same monomeric units units with withinin dimers. dimers. Numbers Numbers on on the the top top of bars of bars mean mean residue numbers, according to the scheme of Figure 1. residueresidue numbers, numbers, according according to to the the scheme scheme of of Figure Figure1 .1. Figure 9 shows that in dimeric eremomycin, there is a large difference in solvent FigureFigure9 9shows shows that that in in dimeric dimeric eremomycin, eremomycin there, there is is a a large large difference difference in in solvent solvent accessibility between N-Ac-D-Ala-filled and free eremomycin dimers. Upon N-Ac-D-Ala accessibilityaccessibility betweenbetweenN N-Ac--Ac-DD-Ala-ﬁlled-Ala-filled andand free free eremomycin eremomycin dimers. dimers. Upon UponN N-Ac--Ac-DD-Ala-Ala binding to eremomycin, water is expelled from the binding pocket, suggesting that the bindingbinding toto eremomycin,eremomycin, waterwater isis expelledexpelled fromfrom thethe binding binding pocket, pocket, suggestingsuggesting thatthat thethe binding site for N-Ac-D-Ala is around the w2, w3 amide groups. In contrast, solvent ac- bindingbinding sitesite forfor N-Ac-D-Ala is is around around the the w2, w2, w3 w3 amide amide groups. groups. In Incontrast, contrast, solvent solvent accessibility did not change far from the binding site. We have shown earlier in similar accessibilitycessibility did did not not change change far far from from the the bind bindinging site. We havehave shownshown earlierearlier inin similarsimilar experiments [27] that the highest saturation is achieved around the binding site. If we experimentsexperiments [[27]27] thatthat thethe highesthighest saturationsaturation isis achievedachieved aroundaround the the binding binding site. site. IfIf wewe compare in straight water saturation transfer experiments the tetramer and octamer comparecompare in in straight straight water water saturation saturation transfer tran experimentssfer experiments the tetramer the tetramer and octamer and mixtureoctamer mixture (Figure 10), we observe differences between them. (Figuremixture 10 (Figure), we observe 10), we differences observe differences between them.between them.

0.8 0.8 brown=tetra brown=tetra blue=octa 0.7 blue=octa 0.7

0.6 0.6

0.5 0.5

0.4 0.4

0.3 0.3

0.2 0.2 Ratioof saturated/unsaturated signals in HSQC

Ratioof saturated/unsaturated0.1 signals in HSQC 0.1

0 0 1 2 3 4 5 6 7 1 2 3NH residue4 number 5 6 7 NH residue number Figure 10. In this "pile up" bar graph, the tetra and octameric eremomycin water saturation effi- FigureFigure 10.10. InIn thisthis “pile"pile up" up” bar bar graph, graph, the the tetra tetra and and octameric octameric eremomycin eremomycin water water saturation saturation effi- efﬁ- ciencies are compared. The smaller the contribution, the higher the probability of solvent access to ciencies are compared. The smaller the contribution, the higher the probability of solvent access to cienciesthat site. are compared. The smaller the contribution, the higher the probability of solvent access to thatthat site. site. Pharmaceuticals 2021, 14, 83 8 of 14

Figure 10 shows qualitatively, that water accessibility is decreased in the octamers (the bottom blue bars) with respect to the tetramers (top brown bars) possibly because of more closed structures, that are already filled with ligands. The w3 amide group around the tetramer binding site has the highest chance for water access. Figure 11. explains the scheme for the arrangement of oligomers in ʺopenʺ and ʺclosedʺ conformations. We were Pharmaceuticals 2021, 14, 83 8 of 14 curious to know how fast the interconversion between the tetra and octamers can hap‐ pen. To learn more about the possible time scale of conversion, EXSY spectra were measuredFigure (Figure 10 shows 12). qualitatively, that water accessibility is decreased in the octamers (the bottom blue bars) with respect to the tetramers (top brown bars) possibly because of Pharmaceuticals 2021, 14, 83 8 of 14 more closed structures, that are already filled with ligands. The w3 amide group around the tetramer binding site has the highest chance for water access. Figure 11. explains the schemeFigure for 10 the shows arrangement qualitatively, of oligomers that water accessibilityin "open" and is decreased"closed" conformations. in the octamers We were (thecurious bottom to blue know bars) how with fast respect the interconversion to the tetramers (top between brown the bars) tetra possibly and becauseoctamers of can hap- morepen. closed To learn structures, more thatabout are alreadythe possible ﬁlled with time ligands. scale Theof conversion, w3 amide group EXSY around spectra were themeasured tetramer (Figure binding site12). has the highest chance for water access. Figure 11. explains the scheme for the arrangement of oligomers in "open" and "closed" conformations. We were curious to know how fast the interconversion between the tetra and octamers can happen. To learn more about the possible time scale of conversion, EXSY spectra were measured (Figure 12).

Figure 11. Schematic arrangement of the oligomers in open or closed conformations. Pink spots represent the ligands that can help to form face‐to‐face interfaces between glycopeptide monomers.

FigureFigure 11. 11.Schematic Schematic arrangement arrangement of the of oligomersthe oligomers in open in op oren closed or closed conformations. conformations. Pink spots Pink spots representrepresent the the ligands ligands that that can helpcan help to form to form face-to-face face-to-face interfaces interfaces between between glycopeptide glycopeptide monomers. monomers. 15N EREMOMYCIN + 1−13C NAc−DADA, 100 ms NOESY

ppm 8.4 15N EREMOMYCIN + 1−13C NAc−DADA, 100 ms NOESY 8.6 ppm 8.8 8.4

9.0 8.6 8.8 9.2 9.0 9.4 9.2

9.6 9.4

9.8 9.6 O2b

10.0 9.8 O2b 10.0 NO EXCHANGE 10.2 NO EXCHANGE 10.2 O2a NOE 2−3 10.4 T2b O2a NOE 2−3 T2a10.4 T2b 10.6 T2a 10.6

10.8 10.8

10.510.5 10.010.0 9.59.5 9.09.0 ppm ppm

Figure 12. EXSY (NOESY) spectrum of 1-13C N-Ac-D-Ala-D-Ala-saturated eremomycin at 278 K, mixingFigure time 12. 100EXS ms.Y (NOESY) The w2 amidespectrum group of signals1-13C N belonging-Ac-D-Ala- toD tetramer-Ala-saturated (T2a, T2b) eremomycin and octamer at 278K, Figure 12. EXSY (NOESY) spectrum of 1‐13C N‐Ac‐D‐Ala‐D‐Ala‐saturated eremomycin at 278K, (O2a,mixing O2b time) are separated100 ms. The on thew2 diagonal,amide group but no signals off-diagonal belonging exchange to tetramer peaks are (T2a, visible T2b) between and octamer mixingtetramer(O2a, time O2b) and 100 octamerare ms. separated The states. w2 However,on amide the diagonal, the group intramolecular but signals no off-diagonal NOESY belonging peaks exchange areto observable.tetramer peaks are(T2a, visible T2b) between and octamer (O2a,tetramer O2b) are and separated octamer states. on the However, diagonal, the intramolecular but no off‐diagonal NOESY peaksexchange are observable. peaks are visible between tetramer and octamer states. However, the intramolecular NOESY peaks are observable.

Pharmaceuticals 2021, 14, 83 9 of 14

The EXSY spectrum (Figure 12) demonstrates the lack of chemical exchange between the octamer and tetramer on the slow NMR timescale (ms-s). Since during titrations no detectable chemical shift movements were observed, we can exclude fast exchange as well. Hence, octamers and tetramers persist simultaneously in solution, without fast exchange of their oligomeric state.

2.3. Results of Antimicrobial Tests 2.3.1. Checkerboard Method The impact of the ligand N-Ac-D-Ala-D-Ala on the glycopeptide activity is shown in Table1.

Table 1. Antibacterial activities of glycopeptides as influenced by the N-Ac-D-Ala-D-Ala dipeptide ligand (checkerboard method).

Concentration of Ligand (µM/mL) Microorganisms Glycopeptide 0 8 16 32 64 128 256 512 1024 Glycopeptide MIC (µM/mL) vancomycin 1 2 2 2 2 2 2 4 8 S. aureus 20450 eremomycin 0.12 0.06 0.06 0.06 0.06 0.125 0.125 0.125 0.5 vancomycin 16 16 16 16 16 16 32 32 64 E. faecalis 9 eremomycin 0.25 0.12 0.25 0.25 0.25 0.25 0.25 0.5 1 vancomycin 1 1 1 1 1 1 2 2 4 S. aureus 209P eremomycin 0.06 0.03 0.03 0.03 0.03 0.03 0.03 0.06 0.125

2.3.2. Disk-Diffusion Method We found that the relative potential depends on the strain selected. For staphylococci, the values were −20% (vancomycin), +20% (eremomycin), regardless of the concentration of glycopeptides and the ligand. Extreme values were found and determined for Enterococcus faecalis 9 (Table2).

Table 2. Antibacterial activities of glycopeptides as inﬂuenced by the N-Ac-D-Ala-D-Ala dipeptide ligand (disk diffusion method) and relative potency against the Enterococcus faecalis 9 strain.

N-Ac-D-Ala-D-Ala (µg/mL) 0 10 100 Glycopeptide Concentration (µg/mL) 5 10 20 5 10 20 5 10 20 diffusion zones (mm) 14.5 16.3 18.1 11.6 13 14.5 8 10.2 12 Vancomycin relative potency - −25% −25% −25% −80% −60% −20% diffusion zones (mm) 15.8 17.6 19.5 17.7 19.4 21.5 18.2 21.1 21.5 Eremomycin relative potency - +12% +10% +10% +15% +20% +10%

3. Discussion Evidence for ligand-induced eremomycin oligomerization has been found in aqueous solution for the ﬁrst time, based on NMR 1H-15N HSQC spectra, chemical shift titration, relaxation, and diffusion measurements. Dimeric eremomycin building blocks oligomer- ize to form discrete octamers and tetramers in the presence of increasing amounts of N-Ac-D-Ala-D-Ala. Dimers were exclusively observed in the presence of similar ligands terminating in the D-Ala sequence. In addition, we demonstrate that increasing ligand Pharmaceuticals 2021, 14, 83 10 of 14

size diminishes or even stops the breathing flip-flop motion of the ring-4 disaccharide that overhangs the binding site, which suggests that exchange due to such motion may happen mainly intramolecularly, without breaking the dimeric structures. It is well known that glycopeptide antibiotics form non-covalent, head-to-tail dimers in aqueous solution with a kind of antiparallel β-strand arrangement of the peptide backbone. In solution NMR, there is a fingerprint indicator of dimerisation, an unusual 1H NMR chemical shift of an aromatic proton signal, due to an orthogonal σ-π interaction of 6e (CH) hydrogen, exhibit- ing a proton chemical shift around 5 ppm, while the 13C shift of 6e-C is in the common aromatic range, at ca. 122 ppm (Figure S1). This effect is always observed in aqueous solutions and is never seen in DMSO solution. Perhaps the lack of water is the reason why dimers were not observed by solid-state NMR, where glycopeptides were investigated using bacterial cell wall lysates, applying trehalose for water replacement. Using X-ray crys- tallography, higher-order oligomers of dimer replicas were often detected and structurally characterized, proving that ligands in the binding sites generate contiguous back-to-back and face-to-face dimers, leading to supramolecular structures. In this work we identified N-Ac-D-Ala-D-Ala-induced eremomycin tetramers and octamers in aqueous solutions, closer to in vivo conditions. In the case of vancomycin, the low dimerization constant and the high affinity to the ligand allow oligomer formations possibly by breaking the dimeric structures. Then, many concurrent oligomers may be formed simultaneously in exchange with each other, hindering NMR observation and characterisation. Still, we attempted to observe if higher-order vancomycin oligomers are formed in aqueous solution upon N-Ac-D-Ala-D-Ala titration. At 280 K, the NMR DOSY experiments displayed a blurred, overlapping diffusion front (not shown), that accords to an average mass around hexamers in the presence of one equivalent ligand. Increasing the temperature to 320 K slightly decreased the average oligomer size. However, it is not yet clear if monomers, dimers, and/or higher-order oligomers of glycopeptide antibiotics persist in vivo, in clinically applied concentrations [34]. Influence of the presence of the cell-wall analogue peptide N-Ac-D-Ala-D-Ala ligand with respect to the antimicrobial activity of glycopeptide antibiotics vancomycin and eremomycin was evaluated by checkboard and disk diffusion methods. The obtained results are presented in Tables1 and2. The results obtained by the checkboard method revealed the following tendency of changes in antimicrobial activity: the decrease of the antibiotic/ligand ratio resulted in a decrease in the MIC values for eremomycin for all strains, while the MIC values for vancomycin in the presence of peptide ligand either did not change or increased (2xMIC) for S. aureus 20450. A significant difference in the interaction of N-Ac-D-Ala-D-Ala and vancomycin or eremomycin was observed in the agar-diffusion assay (Table2, Figure 13). Changes in the concentration of eremomycin and ligand did not affect the relative potential—the zones of inhibition for three strains of test microorganisms increased in the same way as when the antibiotic was combined with N-Ac-D-Ala-D-Ala. When exposed to vancomycin, the effect was microorganism- dependent. For staphylococci, the values were −20%, and as a function of the concentration of vancomycin in combination with the ligand, the diffusion zones decreased proportionally. The extreme values were obtained for Enterococcus faecalis 9. In the presence of 5 µg/mL N-Ac-D-Ala-D-Ala, the zones of inhibition changed the concentration dynamics; the relative potential decreased with increasing antibiotic concentration (Figure 13). PharmaceuticalsPharmaceuticals2021 2021, 14, 14, 83, 83 1111 of of 14 14

25 20 15 10 5 5 µg 10 µg zone of inhibition (mm) 0 20 µg

FigureFigure 13. 13. The change change in in the the diameter diameter zone zone inhibition inhibition in inthe the presence presence antibi antibioticsotics alone alone and andin inin teraction with N-Ac-D-Ala-D-Ala (S. aureus 209P). interaction with N-Ac-D-Ala-D-Ala (S. aureus 209P).

4.4. Materials Materials andand Methods 4.1.4.1.NMR NMR SpectroscopySpectroscopy 1 15 AA Bruker Bruker Avance-II Avance-II NMR NMR spectrometer spectrometer (500.13 (500.13 MHz MHz H1H and and 50.68 50.68 MHz MHz 15NN frequency) frequency) equippedequipped withwith aa multinuclearmultinuclear bbi z-gradient probehead probehead was was used used for for all all measurements measurements ei- eitherther at at 298 298 K or 278278 KK temperature.temperature. DiffusionDiffusion experimentsexperiments werewere carried carried out out using using Bruker’s Bruker’s ‘ledbpgs2s’‘ledbpgs2s’ stimulated stimulated echoecho DOSY DOSY pulse pulse sequence sequence including including bipolar bipolar and and spoil spoil gradients. gradients. ThisThis was was extended extended with with a a 2 2× ×4 4 ms ms spin spin echo echo period period before before the the detection detection period period to to suppress suppress 15 thethe baseline baseline shift shift [ 35[35].]. 15NN relaxationrelaxation experimentsexperiments werewere carriedcarried outout accordingaccording toto [ 29[29]] and and evaluatedevaluated withwith thethe model-freemodel-free approach approach [30,31] [30,31 ]using using the the isotropic isotropic rotational rotational diffusion diffusion ap- approach and a single global correlation time. The recycle delays in T , T experiments proach and a single global correlation time. The recycle delays in T1, T21 experiments2 were were typically 2.5 s, but 5 s in heteronuclear 15N-(1H) NOE. In water saturation differ- typically 2.5 s, but 5 s in heteronuclear 15N-(1H) NOE. In water saturation difference experi- ence experiments, two 15N-1H HSQC experiments were run, using 3 s selective water ments, two 15N-1H HSQC experiments were run, using 3 s selective water presaturation in presaturation in one of them, and the difference was calculated by comparison with an one of them, and the difference was calculated by comparison with an undisturbed reference. undisturbed reference.

4.2.4.2. AntibioticsAntibiotics and Other Reagents Reagents N-Ac-D-Ala-D-Ala was obtained from NovaBioChem (Merck Group, Darmstadt, Ger- N-Ac-D-Ala-D-Ala was obtained from NovaBioChem (Merck Group, Darmstadt, Germany),many), vancomycinvancomycin was wasobtained obtained from fromMerc Merckk (Darmstadt, (Darmstadt, Germany), Germany), while eremomycin while ere- 15 momycinwas prepared was preparedin the Gause in the Institute Gause Instituteof new Antibiotics of new Antibiotics (Moscow, (Moscow, Russia). Russia).N labelled15N ere- la- 13 belledmomycin eremomycin and 1- C and N-Ac- 1-13DC-Ala-N-Ac-D-AlaD-Ala- wereD -Alaproduced were producedas described as describedearlier [27,32]. earlier [27,32].

4.3.4.3. BacterialBacterial StrainsStrains Methicillin-resistantMethicillin-resistant StaphylococcusStaphylococcus aureusaureus(MRSA) (MRSA) strain strain 20450, 20450, vancomycin–resistant vancomycin–resistant EnterococcusEnterococcus faecalis faecalis 9,9,and andStaphylococcus Staphylococcus aureusaureus 209P209P (ATCC(ATCC 6538P) 6538P)were were obtained obtained from from the the MedicalMedical Microbiology Microbiology Laboratory Laboratory of of State State Research Research Center Center for for Antibiotics Antibiotics ((Moscow,Moscow, RussiaRussia).).

4.4.4.4. DeterminationDetermination of Antibiotic Minimum Minimum Inhibitory Inhibitory Concentration Concentration EvaluationEvaluation ofof MinimumMinimum Inhibitory Concentrations Concentrations (MIC, (MIC, µg/mL)µg/mL) of ofvancomycin vancomycin and anderemomycin eremomycin was wascarried carried out using out usingthe microdilution the microdilution method methodin Mueller–Hinton in Mueller–Hinton broth. For broth.MIC determination, For MIC determination, 96-well microtiter 96-well microtiterplates, cont plates,aining containing a dilution aseries dilution of the series antibiotics, of the 5 ◦ antibiotics,were inoculated were inoculatedwith 105 CFU/mL with 10 andCFU/mL incubated and at 37 incubated °C for 24at h. 37 C for 24 h. AA range range of of antibiotic antibiotic concentrations concentrations based based on on the the obtained obtained MIC MIC values values was was selected selected to to analyzeanalyze thethe interactioninteraction of glycopeptides with with NN-Ac--Ac-DD-Ala--Ala-D-Ala.D-Ala. Pharmaceuticals 2021, 14, 83 12 of 14

4.5. Interaction of Glycopeptides and the Ligand 4.5.1. Checkerboard Method A range of antibiotic concentrations based on the obtained MIC values was selected to analyze the interaction of glycopeptides with N-Ac-D-Ala-D-Ala. Interactions between glycopeptides and dipeptide were investigated in 96-well microtiter plates in the Mueller– Hinton broth. The inoculum contained 105 CFU/mL. The MIC of the glycopeptides (µM) for each concentration of ligand was read 18 h after incubation at 37 ◦C. Microtiter plates with the antibiotic and peptide ligand were prepared either in ad- vance and kept at 4 ◦C for 18 h or prepared immediately before inoculum introduction.

4.5.2. Disk-Diffusion Method The standard 6 mm discs to which 20 µL of a solution containing 10 or 100 µg of N-Ac-D-Ala-D-Ala in H2O (0.5 or 5 mg/mL) was applied were dried in a laminar ﬂow at room temperature. Then, the solution of the antibiotic was added and the disks were dried again. Glycopeptide concentrations were 5, 10, and 20 µg per disc. The disks were placed on agar medium, containing 108 CFU/mL of microorganism strain. The discs loaded with glycopeptide antibiotics solely were used to compare the zones of inhibition. Zones of inhibition were measured after 18–24 h of incubation at 37 ◦C. The diameters of the zones of inhibition were measured with a caliper, and the average of nine values for each experimental point was calculated. The relative potential was determined by dividing the larger diameter by the smaller one. The sign (–) corresponds to a decrease, and a positive value to an increase in the zone of inhibition of the antibiotic N-Ac-D-Ala-D-Ala.

5. Conclusions Higher-order glycopeptide antibiotic oligomers, first suggested by crystal studies, are here shown to be formed in aqueous solution as well. However, in contrast to X-ray-detected vancomycin hexamers [18], we detected a mixture of tetra-and octameric eremomycin in solution, depending on the pertinent dipeptide ligand concentration. Interestingly, oligomerisation was induced exclusively by the dipeptide N-Ac-D-Ala-D-Ala. Neither N-Ac-D-Ala nor the cell wall analogue tripeptide (N-Ac)2-Lys-D-Ala-D-Ala induced higher-order oligomerisation. In any glycopeptide antibiotic, oligomeric-form NMR signal doubling is observed, characteristic of asymmetric back- to back dimer building blocks contributing to the pseudo- C2 symmetry of higher-order oligomers. It is not yet clear if such oligomers are also generated in vivo when glycopeptides are used for medical treatments, though slight increases in antibacterial effect could be induced by some model peptides. Synthetic, multivalent ligands provide a promising alternative to enhance ligand binding [36]. Ligand-induced oligomer formation of eremomycin slightly increased the antibacterial activity when the ligand concentration was low at the beginning, but at increased N-Ac-D-Ala-D-Ala concentrations, the competition reduced this beneficial effect. If eremomycin was loaded at a stoichiometrically relevant ligand quantity from the starting point, then adding more ligand in consecutive steps did not change the antimicrobial activity. The differences in antibacterial effect between vancomycin and eremomycin in the presence of N-Ac-D-Ala-D-Ala may result from the differences in their mode of action affecting transpeptidation and transglycosylation. Specifically, ligand-induced and -mediated reversible oligomerization of glycopeptide antibiotics is an interesting example of supramolecular structures, and formation of ligand- controlled peptide self-assemblies is challenging in their own right, which might even have a distant analogy with amyloid plaque formation [37]. Reprogramming or modulating the antimicrobial effects in vancomycin-type antibiotics with covalent modifications might be promising, since some glycopeptides were shown to have antiviral effects [38]. Pharmaceuticals 2021, 14, 83 13 of 14

Supplementary Materials: The following are available online at https://www.mdpi.com/1424-824 7/14/2/83/s1, Figure S1: 1H-13C HSQC spectrum of the eremomycin dimer at 298 K, at pH 4.5 in 20 mM acetate (D2O) buffer (Bruker NEO-700 NMR spectrometer, equipped with Prodigy, TCI probehead). Materials and methods for antimicrobial tests with N-Ac-D-Ala-D-Ala- loaded eremomycin. Author Contributions: Conceptualization, G.B. and P.H.; methodology L.I. and G.B.; software, G.B. and L.I.; validation, G.B., L.I. and A.N.T.; NMR investigation, L.I., R.E. and G.B.; microbiology investigation, A.N.T., N.E.G. and A.E.S.; writing—original draft preparation, G.B., L.I.; writing—review and editing, G.B., L.I., R.E., P.H., A.E.S. and A.N.T.; visualization, L.I., R.E. and G.B.; supervision G.B., A.N.T. and P.H., funding acquisition, G.B. an P.H. All authors have read and agreed to the published version of the manuscript. Funding: This research was funded by the EU and co-financed by the European Regional Develop- ment Fund under the projects GINOP-2.3.2-15-2016-00008 to G.B. and GINOP-2.3.3-15-2016-00004 (access to Bruker NEO 700 MHz NMR spectrometer). The project was supported by the National Research and Development and Innovation Office of Hungary K119509 (G.B. and P.H.). Institutional Review Board Statement: Not applicable. Informed Consent Statement: Not applicable. Data Availability Statement: More data of this work are available in supplementary material. Acknowledgments: We are grateful to M.N. Preobrazhenskaya (passed away on 25 December 2014) and Patrick Groves for ideological support of early work in the glycopeptide antibiotics project. Conflicts of Interest: The authors declare no conflict of interest.

References 1. Healy, V.L.; Lessard, I.A.D.; Roper, D.I.; Knox, J.R.; Walsh, C.T. Vancomycin resistance in enterococci: Reprogramming of the D-Ala-D-Ala ligases in bacterial peptidoglycan biosynthesis. Chem. Biol. 2000, 7, R109–R119. [CrossRef] 2. Reynolds, P.E.; Courvalin, P. Vancomycin resistance in enterococci due to synthesis of precursors terminating in D-Alanyl-D-Serine. Antimicrob. Agents Chemother. 2005, 49, 21–25. [CrossRef][PubMed] 3. Zeng, D.; Debabov, D.; Hartsell, T.L.; Cano, R.J.; Adams, S.; Schuyler, J.A.; McMillan, R.; Pace, J.L. Approved Glycopeptide Antibacterial Drugs: Mechanism of Action and Resistance. Cold Spring Harb. Perspect. Med. 2016, 6, a026989. [CrossRef][PubMed] 4. Blaskovich, M.A.T.; Hansford, K.A.; Butler, M.S.; Jia, Z.G.; Mark, A.E.; Cooper, M.A. Developments in Glycopeptide Antibiotics. Acs Infect. Dis. 2018, 4, 715–735. [CrossRef][PubMed] 5. Loll, P.J.; Axelsen, P.H. The structural biology of molecular recognition by vancomycin. Annu. Rev. Biophys. Biomol. Struct. 2000, 29, 265–289. [CrossRef] 6. Waltho, J.P.; Williams, D.H. Aspects of Molecular Recognition - Solvent Exclusion and Dimerization of the Antibiotic Ristocetin When Bound to a Model Bacterial Cell-Wall Precursor. J. Am. Chem. Soc. 1989, 111, 2475–2480. [CrossRef] 7. Batta, G.; Sztaricskai, F.; Kover, K.E.; Rudel, C.; Berdnikova, T.F. An NMR Study of Eremomycin and its Derivatives Full H-1 and C-13 Assignment, Motional Behavior, Dimerization and Complexation with Ac-D-ALA-D-ALA. J. Antibiot. 1991, 44, 1208–1221. [CrossRef] 8. Gerhard, U.; Mackay, J.P.; Maplestone, R.A.; Williams, D.H. The Role of the Sugar and Chlorine Substituents in the Dimerization of Vancomycin Antibiotics. J. Am. Chem. Soc. 1993, 115, 232–237. [CrossRef] 9. Groves, P.; Searle, M.S.; Mackay, J.P.; Williams, D.H. The Structure of an Asymmetric Dimer Relevant to the Mode of Action of the Glycopeptide Antibiotics. Structure 1994, 2, 747–754. [CrossRef] 10. Beauregard, D.A.; Williams, D.H.; Gwynn, M.N.; Knowles, D.J.C. Dimerization and Membrane Anchors in Extracellular Targeting of Vancomycin Group Antibiotics. Antimicrob. Agents Chemother. 1995, 39, 781–785. [CrossRef] 11. Batta, G.; Cristofaro, M.F.; Sharman, G.J.; Williams, D.H. Demonstration of the difference in binding affinity between the two binding sites of the ristocetin A asymmetric dimer. Chem. Commun. 1996, 1, 101–103. [CrossRef] 12. Bardsley, B.; Williams, D.H. Measurement of the different affinities of the two halves of glycopeptide dimers for acetate. Chem. Commun. 1997, 11, 1049–1050. [CrossRef] 13. Shiozawa, H.; Chia, B.C.S.; Davies, N.L.; Zerella, R.; Williams, D.H. Cooperative binding interactions of glycopeptide antibiotics. J. Am. Chem. Soc. 2002, 124, 3914–3919. [CrossRef][PubMed] 14. McPhail, D.; Cooper, A. Thermodynamics and kinetics of dissociation of ligand-induced dimers of vancomycin antibiotics. J. Chem. Soc.-Faraday Trans. 1997, 93, 2283–2289. [CrossRef] 15. Linsdell, H.; Toiron, C.; Bruix, M.; Rivas, G.; Menendez, M. Dimerization of A82846B, vancomycin and ristocetin: Influence on antibiotic complexation with cell wall model peptides. J. Antibiot. 1996, 49, 181–193. [CrossRef] Pharmaceuticals 2021, 14, 83 14 of 14

16. Loll, P.J.; Miller, R.; Weeks, C.M.; Axelsen, P.H. A ligand-mediated dimerization mode for vancomycin. Chem. Biol. 1998, 5, 293–298. [CrossRef] 17. Schafer, M.; Sheldrick, G.M.; Schneider, T.R.; Vertesy, L. Structure of balhimycin and its complex with solvent molecules. Acta Crystallogr. Sect. D Biol. Crystallogr. 1998, 54, 175–183. [CrossRef] 18. Loll, P.J.; Derhovanessian, A.; Shapovalov, M.V.; Kaplan, J.; Yang, L.; Axelsen, P.H. Vancomycin Forms Ligand-Mediated Supramolecular Complexes. J. Mol. Biol. 2009, 385, 200–211. [CrossRef] 19. Jia, Z.G.; O’Mara, M.L.; Zuegg, J.; Cooper, M.A.; Mark, A.E. Vancomycin: Ligand recognition, dimerization and super-complex formation. FEBS J. 2013, 280, 1294–1307. [CrossRef] 20. Loll, P.J.; Bevivino, A.E.; Korty, B.D.; Axelsen, P.H. Simultaneous recognition of a carboxylate-containing ligand and an intramolecular surrogate ligand in the crystal structure of an asymmetric vancomycin dimer. J. Am. Chem. Soc. 1997, 119, 1516–1522. [CrossRef] 21. Lehmann, C.; Bunkoczi, G.; Vertesy, L.; Sheldrick, G.M. Structures of glycopeptide antibiotics with peptides that model bacterial cell-wall precursors. J. Mol. Biol. 2002, 318, 723–732. [CrossRef] 22. Cooper, M.A.; Williams, D.H. Binding of glycopeptide antibiotics to a model of a vancomycin-resistant bacterium. Chem. Biol. 1999, 6, 891–899. [CrossRef] 23. Cooper, M.A.; Williams, D.H.; Cho, Y.R. Surface plasmon resonance analysis of glycopeptide antibiotic activity at a model membrane surface. Chem. Commun. 1997, 17, 1625–1626. [CrossRef] 24. Good, V.M.; Gwynn, M.N.; Knowles, D.J.C. MM-45289, A Potent Glycopeptide Antibiotic which Interacts Weakly with Diacetyl- L-Lysyl-D-Alanyl-D-Alanine. J. Antibiot. 1990, 43, 550–555. [CrossRef][PubMed] 25. Beauregard, D.A.; Maguire, A.J.; Williams, D.H.; Reynolds, P.E. Semiquantitation of cooperativity in binding of vancomycin-group antibiotics to vancomycin-susceptible and -resistant organisms. Antimicrob. Agents Chemother. 1997, 41, 2418–2423. [CrossRef] 26. Mackay, J.P.; Gerhard, U.; Beauregard, D.A.; Westwell, M.S.; Searle, M.S.; Williams, D.H. Glycopeptide Antibiotic-Activity and the Possible Role of Dimerization—A Model for Biological Signaling. J. Am. Chem. Soc. 1994, 116, 4581–4590. [CrossRef] 27. Batta, G.; Sztaricskai, F.; Makarova, M.O.; Gladkikh, E.G.; Pogozheva, V.V.; Berdnikova, T.F. Backbone dynamics and amide proton exchange at the two sides of the eremomycin dimer by N-15 NMR. Chem. Commun. 2001, 5, 501–502. [CrossRef] 28. Allen, N.E.; Nicas, T.I. Mechanism of action of oritavancin and related glycopeptide antibiotics. FEMS Microbiol. Rev. 2003, 26, 511–532. [CrossRef] 29. Farrow, N.A.; Muhandiram, R.; Singer, A.U.; Pascal, S.M.; Kay, C.M.; Gish, G.; Shoelson, S.E.; Pawson, T.; Formankay, J.D.; Kay, L.E. Backbone Dynamics of a Free and a Phosphopeptide -Complexed SRC Homology-2 Domain Studied by N-15 NMR Relaxation. Biochemistry 1994, 33, 5984–6003. [CrossRef] 30. Lipari, G.; Szabo, A. Model-Free Approach to the Interpretation of Nuclear Magnetic-Resonance Relaxation in Macromolecules. 1. Theory and Range of Validity. J. Am. Chem. Soc. 1982, 104, 4546–4559. [CrossRef] 31. Lipari, G.; Szabo, A. Model-Free Approach to the Interpretation of Nuclear Magnetic-Resonance Relaxation in Macromolecules. 2. Analysis of Experimental Results. J. Am. Chem. Soc. 1982, 104, 4559–4570. [CrossRef] 32. Batta, G.; Kover, K.E.; Szekely, Z.; Sztaricskai, F. Glycopeptide Binding-Site Spied through Transferred Heteronuclear NOE - 1-C-13 Ac-D-Ala-D-Ala Bonded to Vancomycin and Ristocetin-A. J. Am. Chem. Soc. 1992, 114, 2757–2758. [CrossRef] 33. Sharman, G.J.; Searle, M.S.; Benhamu, B.; Groves, P.; Williams, D.H. Burial of Hydrocarbon Causes Cooperative Enhancement of Electrostatic Binding. Angew. Chem. Int. Ed. Eng. 1995, 34, 1483–1485. [CrossRef] 34. Phillips-Jones, M.K.; Lithgo, R.; Dinu, V.; Gillis, R.B.; Harding, J.E.; Adams, G.G.; Harding, S.E. Full hydrodynamic reversibility of the weak dimerization of vancomycin and elucidation of its interaction with VanS monomers at clinical concentration. Sci. Rep. 2017, 7, 12697. [CrossRef][PubMed] 35. Orosz, L.; Batta, G.; Keki, S.; Nagy, M.; Deak, G.; Zsuga, M. Self-association of bis-(alpha,beta-D-glucopyranosyl)-polyisobutylene. Carbohydr. Res. 2007, 342, 1323–1328. [CrossRef][PubMed] 36. Nicolaou, K.C.; Boddy, C.N.C.; Brase, S.; Winssinger, N. Chemistry, biology, and medicine of the glycopeptide antibiotics. Angew. Chem.-Int. Ed. 1999, 38, 2096–2152. [CrossRef] 37. Iadanza, M.G.; Jackson, M.P.; Hewitt, E.W.; Ranson, N.A.; Radford, S.E. A new era for understanding amyloid structures and disease. Nat. Rev. Mol. Cell Biol. 2018, 19, 755–773. [CrossRef][PubMed] 38. Sz˝ucs,Z.; Naesens, L.; Stevaert, A.; Ostorházi, E.; Batta, G.; Herczegh, P.; Borbás, A. Reprogramming of the Antibacterial Drug Vancomycin Results in Potent Antiviral Agents Devoid of Antibacterial Activity. Pharmaceuticals 2020, 13, 139. [CrossRef]