University of Groningen Nonribosomal peptide synthetases Zwahlen, Reto Daniel

(1)

University of Groningen

Nonribosomal peptide synthetases

Zwahlen, Reto Daniel

IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below.

Document Version

Publisher's PDF, also known as Version of record

Publication date: 2018

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA):

Zwahlen, R. D. (2018). Nonribosomal peptide synthetases: Engineering, characterization and biotechnological potential. University of Groningen.

Copyright

Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).

Take-down policy

If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.

(2)

CHAPTER IV

Biochemical

and structural

characterization of the

Nocardia lactamdurans

L-δ-(α-aminoadipyl)-L-cysteinyl-D-valine

synthetase

Reto D. Zwahlen,1 Riccardo Iacovelli,1 Sathish N. Yadav Kadapalakere,2

Gert T. Oostergetel,2 Roel A.L. Bovenberg,3,4 and Arnold J.M. Driessen1,5

1Molecular Microbiology, Groningen Biomolecular Sciences and Biotechnology Institute, University of Groningen, Groningen, The Netherlands

2Electron Microscopy, Groningen Biomolecular Sciences and Biotechnology Institute, University of Groningen, Groningen, The Netherlands

3Synthetic Biology and Cell Engineering, Groningen

Biomolecular Sciences and Biotechnology Institute, University of Groningen, Groningen, The Netherlands

4DSM Biotechnology Centre, Delft, The Netherlands 5Kluyver Centre for Genomics of Industrial Fermentations,

(3)

Abstract

The L-δ-(α-aminoadipyl)-L-cysteinyl-D-valine synthetase (ACVS) is a nonribosomal peptide synthetase (NRPS) that fulfils a crucial role in the biosynthesis of β-lactams. Although some of the en-zymological aspects of ACVS have been elucidated, the large size of the protein, over 400 kD, has hampered expression and stable purification. To biochemically and structurally characterize the enzyme, the Nocardia lactamdurans ACVS was expressed in E. coli HM0079. Using a two-step, affinity — size exclusion approach, the protein was purified to homogeneity and characterized for peptide formation. Conditions were established allowing for the stabiliza-tion of the flexible ACVS protein and electron microscopic analysis yielding a first low resolution structural model for a multi-modular NRPS.

(4)

Biochemical and structur al char acterization o f the Noc ar dia lactamdur ans L-δ-( α-aminoadip yl )-L -cy stein yl -D-valine s ynthetase

4 Introduction

Nonribosomal peptides (NRP) represent a versatile group of low to medium molecular weight compounds that exhibit various biological activities. These peptides are exclusively produced by nonribosomal peptide synthetases (NRPS) and do not only contain proteinogenic amino acids, but may also con-tain a wide variety of non-proteinogenic amino acids or carboxylic acids [1]. NRP often undergo a series of modifications in trans, whether through the action of the NRPS or by further tailoring enzymes. Due to the relative sim-plicity and overall significance, the b-lactam production pathway has been a paradigm for related research fields. Three distinct enzymatic steps are involved in the production of β-lactam, with the NRPS L-δ-(α- aminoadipyl)-L-cysteinyl-D-valine synthetase (ACVS) providing the tripeptide L-δ-(α- aminoadipyl)-L-cysteinyl-D-valine (LLD-ACV) as the precursor for β-lactam antibiotics such as penicillins or cephalosporines (Figure 1). ACVS is a three modular NRPS responsible for the incorporation of L-α-aminoadipic acid, L-cysteine and L-valine into the LLD-ACV tripeptide. These amino acids are inserted in the final product in a co-linear fashion, thus the position of the incorporated substrate corresponds to the position of the respective module within the primary NRPS sequence [2]. Peptide formation itself is strictly determined by the selectivity of the domains of the ACVS. NRP synthesis universally starts in every module, with the adenylation (A) domain, serving as a highly selective gate keeper, which recruits and adenylates a distinct substrate, thereby forming an adenyl-substrate conjugate. Subsequently, the conjugate is transferred to the downstream thiolation (T) domain, which serves in its active, phosphopantheteinated (ppant) state as a flexible linker domain, guiding the activated substrates to the donor- and acceptor sites of the up- or downstream condensation (C) domains, where peptide formation occurs with the upstream substrate being released from the ppant moiety. Ultimately, the newly synthesized peptide is modified either in cis or in trans. In ACVS, L-valine is epimerized via an intrinsic epimerization (E) domain, and finally released in a thioesterification reaction by means of a highly selec-tive thioesterase (Te) domain. LLD-ACV production takes place in the cytosol, however downstream processing of the tripeptide is compartmentalized in filamentous fungi where the cytosolic isopenicillin-N synthase (IPNS) forms the β-lactam ring, and the microbody localized acyltransferase replaces the aminoadipate moiety of isopenicillin N (IPN) for an alternative side chain. Due to the importance of the β-lactam biosynthetic pathway and wide spread of the ACVS across a range of organisms [3–4], it has been the focus point of functional studies. ACVS served as a model NRPS, aiding to establish

(5)

92

4

Intr oduction E A T A Te C A D A S O T T S O S O IPNS C C _DA O S H N H2 C H3 C H3 O N H2 H N H2 O O H O N H O H O O N H2 O S H N H C H3 C H3 O H O N H O H O O N H2 O S N C H3 C H3 O H O N O O H S C H3 N H O H N H2 O O C H3 N O O H S C H3 N H N H2 O O C H3 R 1 N H O O S N O_H O R 2 H R 1 N H O O S N C H3 C H3 O H O Semi-s yn the �c P enicillins Cephalisporines Penicillins

(6)

4

the NRPS thiotemplate mechanism [5], and has since been studied in fungi such as penicillium, aspergilli, cephalosporium as well as bacterial nocardia and

streptomyces species [6–12]. The biochemical reactions of NRPS sub-domains

has been elucidated [13–14], however, studies on the isolated and purified ACVS are scarce [4]. The large size of the protein, 404–425 kD [4], makes it a challenge with respect to expression and purification, to gain appropri-ate amounts of sufficient quality and quantity. In combination with the high conformational dynamics and flexibility that characterizes NRPS enzymes, structural analysis has only progressed slowly [15]. Despite extensive efforts, including the solution of sub-domain, domain, di-domain and entire mod-ular structures [16], no reliable structure of an entire multi-modmod-ular NRPS has yet been solved. On the basis of available crystal structures of (sub-) domains and the determined interfaces, a model for multi-modular NRPS enzymes has been proposed [17]. Pilot studies using a combinatorial, crystal-lographic and electron microscopic approach, underlined the validity of the proposed model and further illustrate the additional complexity, originating in the highly dynamic conformation of the NRPS enzymes [18]. Here, we fo-cus on the pcbAB gene of the organism Nocardia lactamdurans that encodes a 404 kD ACVS [12]. To characterize this enzyme, we expressed the protein us-ing the E. coli strain HM0079 [19] as a platform and subsequently purified the enzyme to homogeneity. This allowed for the determination of fundamental biochemical parameters and the intrinsic substrate specificity of the enzyme. Through stabilization by co-factors, substrates and gradient fixation (Grafix) [20] conditions, a conformationally homogenous sample was obtained that for the first time yielded a low resolution structure using negative stain elec-tron microscopy.

Figure 1 — ACVS domain organization and product formation.

The ACVS consists of a total of 10 domains arranged in three modules with distinct speci-ficities for the incorporation of L-α-aminoadipic acid (L-α-aaa), L-cysteine and L-valine into the tripeptide δ-(L-α-aminoadipoyl)-L-cysteinyl-D-valine (LLD-aaa-cys-val). The domain ar-rangement is conserved [4] and follows the order: N-1(AT)-2(CAT)-3(CATTe)-C. The resulting LLD-aaa-cys-val is converted into isopenicillin-N (IPN) guided by the isopenicillin-N synthe-tase (IPNS). In the penicillin biosynthetic pathway, IPN is further converted into a distinct penicillin compound dependent upon the available side chain. Other routes can result in the formation of cephalosporines and related compounds [36], while semi-synthetic routes are applied to produce ampicillin or amoxicillin.

(7)

94

4

Material and methods

Strains, plasmids and general culturing conditions

All cloning procedures were performed using E. coli DH5α and cultures were grown using LB medium at 37 °C and 200 rpm and antibiotic selection was conducted utilizing 25 µg/ml zeocin. The Nocardia lactamdurans pcbAB was cloned using an intermediate gateway vector and was subsequently sub-cloned into the pBAD-plasmid (pBR322 ori; araC; pBAD, ZEO) using SbfI × NdeI sites including the introduction of a 6×his tag on the C-terminal end. This con-struct was kindly provided by DSM Sinochem Pharmaceuticals BV.

Expression and his-tag affinity purification of ACVS

Cultures were grown to an OD600 of 0.6, transferred to 18 °C and 200 rpm for 1 h and subsequently induced using 0.3 mM IPTG and 0.2 % L-arabinose. Harvest was done 18 h after induction by spinning at 3500 g for 15 minutes. After resuspension in lysis buffer (50 mM HEPES pH 7.0, 300 mM NaCl, 2 mM DTT, Complete EDTA free protease inhibitor; Roche No. 04693159001), cells were disrupted using sonication (6 s/15 s; on/off, 50x, 10 µm amplitude) and cell-free lysate obtained by centrifugation at 4 °C, 13000 g, 15 minutes. Pu-rified enzyme was extracted by means of Ni-NTA bead (Qiagen) supported his-tag affinity purification using gravity flow. Wash steps were performed using two column volumes of wash buffer (50 mM HEPES pH 7.0, 300 mM NaCl, 20 mM imidazole) followed by a three-step elution using one bed vol-ume of each elution buffer (50 mM HEPES pH 7.0, NaCl 300 mM, imidazole 50/150 or 250 mM). Samples were concentrated if necessary, using Amicon U-100 spin filters (Amicon). Final concentration was determined using A280 or the DC protein assay (Biorad).

In vitro product formation assay

Isolated enzymes were subjected to in vitro assays, in order to determine product formation kinetics. Assay conditions initially used include 50 mM HEPES pH 7.0, 300 mM NaCl, 5 mM ATP pH 7.0, 100 µM CoA, 0.2 µM phospho-pantetheinyl transferase (Sfp, NEB), 5 mM L-α-aminoadipic acid (aaa), 2 mM L-cysteine (cys), 2 mM L-valine (val), 5 mM MgCl2, 2 mM DTT and 0.17 µM ACVS. For velocity and affinity determination, amino acid concentration of

(8)

4

0.1; 0.25; 0.5; 1; 2 and 5 mM were used, for specificity determination the concentration of the variable amino acid was set at 5 mM. Reactions were run at 30 °C and sampling took place after 0, 10, 20, 30, 45, 60, 120 and 240 minutes for dynamic measurements and after 0 and 240 or 960 min-utes for endpoint value determination. 0.1 M NaOH was used to stop the reactions. Samples were subsequently stored at −80 °C and reduced before LC/MS analysis using 10 mM DTT or TCEP.

Liquid chromatographic and masspectrometric analysis (LC/MS)

Samples (50 µl) obtained from an in vitro reaction were subjected to LC/MS analysis. Two technical replicates were run per sample at 5 µl each. Analy-sis was performed using a LC/MS Orbitrap (Thermo Scientific) in combination with a RP-C18 column (Shimadzu Shim pack XR-ODS 2.2; 3.0 × 75 mm). Scan range was set at 80–1600 M/Z in positive Ion (4.2 kV spray, 87.5 V capillary and 120 V of tube lens) mode, with capillary temperature set at 325 °C. A gra-dient program with miliQ water (A), acetonitrile (B) and 2 % formic acid (C) was run; 0 min; A 90 %, B 5 %, C 5 %; 4 min, A 90 %, B 5 %, C 5 %; 13 min, A 0 %, B 95 %, C 5 %; 16 min A 0 %, B 95 %, C 5 %; 16 min, A 90 %, B 5 %, C 5 %; 21 min A 90 %, B 5 %, C 5 % at a flow rate of 0.3 ml min-1. The Bis-aaa-cys-val standard was obtained from Bachem and used for quantification in a standard curve at concentrations of 0.1; 0.5; 1; 5; 10; 50 and 100 µM. Novel tripeptides were iden-tified according to accurate monoisotopic mass, if not mentioned otherwise.

Affinity and Gel-filtration chromatography

His-tag affinity purified samples were utilized as a basis for further refining using a FPLC (Amersham, FPLC) linked system in combination with a XD-16 column (Amersham, L × D, volume 65 ml) and Superdex 200 column bed material (Superdex 200, Sigma Aldrich). The system was equilibrated using miliQ first, at a flow rate of 1 ml/min for 1.5 h, followed by sample buffer equilibration (50 mM HEPES pH 7.0, 300 mM NaCl) at the same flow rate for 2 h. Thereafter, 500–1000 µl of sample (Elution fraction or concentrate) was injected (1 ml loop), run over the equilibrated column at a flow rate of 0.2–0.5 ml/min and fractions of 0.5–1 ml were collected. The obtained pro-tein fractions were subsequently measured at either A280 or using the DC protein assay (BioRad) in order to determine the protein concentration and thereafter all collected fractions were visualized on a 5 % SDS-PAGE, 17.5 %

(9)

96

4

Material and methods

Tricine gel or a precast 5–20 % gradient gel (BioRad Nupage, Biorad) to pin-point the fraction with the highest purity and specific protein concentra-tion. Purity was determined using 2-D densitometry and gel-images were processed on AIDA software (AIDA image analyzer). The peak fractions were furthermore analyzed using dynamic light scattering (DLS) to determine par-ticle size and distribution as a parameter for sample homogeneity. DLS was conducted on a DynaPro Nanostar (Wyatt) at 4 °C and 80 % laser intensity.

Grafix gradient separation of purified ACVS

In parallel to the FPLC based purification methods, a gradient centrifugation method was applied, derived from the GraFix method. Therefore, a glycerol gra-dient in the range from 10–30 % was utilized, supplemented with 50 mM HEPES pH 7.0 and 300 mM NaCl. Gradients were created using a series of glycerol-buf-fer solutions (10; 12.5; 15; 17.5; 20; 22.5; 25; 27.5 and 30 % glycerol), layered in an ultracentrifugation tube and frozen in liquid nitrogen after the addition of every solution. The cross-linking agent glutaraldehyde (GA) was applied to the gradi-ent, therefore 0.05–0.2 % were added to the 30 % glycerol- buffer solution. Fro-zen gradients were subsequently thawed on ice and 100–1500 pmol of his-tag purified ACVS was carefully added to the top. The tubes were thereafter centri-fuged at 250000 g for 60–90 minutes, removed and fractionated, using a per-istaltic pump in combination with an 110 × 0.8 mm needle at 0.1–0.2 ml/min. The obtained 250–500 µl fractions were analyzed on 5 % SDS-PAGE, 17.5 %

tri-cine gel or a precast 5–20 % gradient gel (BioRad Nupage, Biorad) as well as us-ing A280 or the DC protein assay (biorad) to measure the concentration. Selected fractions were ultimately subjected to DLS and transferred to the EM facilities for negative stain analysis.

Electron-microscopic (EM) and cryo-EM structural analysis of ACVS

particles

Negatively stained specimens for transmission electron microscopy were ppared on carbon coated copper grids using 2 % uranyl acetate. Images were re-corded with a Tecnai G2 20 Twin transmission electron microscope (FEI, Eind-hoven, the Netherlands), operated at 200 kV and equipped with an UltraScan 4000UHS CCD camera (Gatan, Pleasanton, CA, USA) with a pixel size of 2.24 Å at the specimen level after binning the images to 2048 × 2048 pixels. For cryo-electron microscopy 3 µl aliquots of the sample were applied onto

(10)

4

glow-discharged holey carbon grids (Quantifoil R2/2). Grids were blotted for 5 s at 100 % humidity, before being vitrified in a FEI Vitrobot using liq-uid ethane. Grids were transferred to a G2-Polara (FEI), operating at 300 kV, equipped with a Gatan imaging energy filter (GIF) in zero-loss mode and a Gatan 2K CCD camera. Micrographs were recorded with a pixel size of 2.63 Å at the specimen level, with a defocus of 2 µm and with a dose of 25 e− Å−2. Electron micrographs were processed using Relion-1.3 software [21]. The con-trast transfer function parameters of the G2-Polara micrographs were deter-mined with CTFFIND3 [22].

Results

Expression, purification and biochemical characterization of the

Nocardia lactamdurans ACVS

The gene encoding the Nocardia lactamdurans ACVS was overexpressed in E.

coli HM0079 as a C-terminal 6×his-tagged protein, and purified by Ni+

affin-ity purification followed by size exclusion chromatography. The overall yield from shaking cultures was 13.9 ± 3.4 mg pure ACVS per liter of culture. The isolate of the first step affinity purification lead to a protein solution with a purity of 32.1 ± 4.8 %, which increased to 90.2 ± 0.5 after size exclusion chromatography using gel filtration by FPLC (Figure 2). The purified ACVS was subjected to in vitro product formation assays using conditions outlined in the methods section and in supplementary figure 1. A set of distinct assay variations were prepared using varying concentrations of the three substrate amino acids in combination with 0.17 µM ACVS (Figure 3). Reactions were evaluated over a 4 h time course, and analyzed on LC/MS. Resulting LLD-aaa-cys-val levels were quantified and normalized, showing near to linear prod-uct formation curves (Figure 3A–C) over the indicated time course. Maximal LLD-ACV product levels under the given conditions reached roughly 50 µM after an assay time of 4 h which exceeds the enzyme concentration by more than two orders of magnitude indicating multiple turnovers. The enzyme catalytic rate was determined and expressed as product formed per minute and µM ACVS. The calculated Kcat value for the ACVS was 0.78 ± 0.14 min−1. KM values were determined from the Michalis mention kinetics with a >98 % curve fit. Values of 640 ± 16, 40 ± 1 and 150 ± 4 µM were determined for L-α-aminoadipic acid, L-cysteine and L-valine respectively (Figure 3D).

(11)

98

4

Results

Figure 2 — Two step purification of the Nocardia lactamdurans ACVS.

(A) ACVS was isolated from E. coli HM0079 cells and harvested after overnight expression at

18 °C. A cell free lysate was obtained through sonication and subsequently separated into a clear supernatant (CFL) and the pellet was resuspended in 8 M UREA (CFL (i)). The clear lysate was further purified using gravity flow in combination with a his-tag affinity chroma-tography, using two washing steps (W1, W2) and elution with 50, 150 and 250 mM imidazole, respectively (E1, E2, E3). (B) Fractions E1 and E2 were subjected to gel filtration on FPLC. 1 ml of an elution fraction was injected and fractions were collected in the same volume (1–14). Fractions 10–12 were further analyzed using negative stain EM, cryo EM and biochemical characterization methods. ACVS yields were 13.9±3.4 mg per liter of culture and the purity sets at 32.1 ± 4.8 % and 90.2 ± 0.5 % for samples obtained from the first and second purifi-cation, respectively.

▽

= single particle;

▼

= aggregation.

(12)

4 Substrate specificity of the N. lactamdurans ACVS

Next we determined the enzyme substrate specificity. Therefore, three sets of reactions were arranged, varying the substrate for each of the three ACVS modules, using structurally analogous substrates. The concentration of the variable amino acid was set at 5 mM. Product levels were determined as end points after 4 h and analyzed for the formation of the predicted tripep-tides and related structures using LC/MS. In addition to the three native sub-strates, 14 analogues were tested in a total of 22 reaction setups (Figure 4). Next to the LLD-ACV tripeptide, we managed to detect 10 of the proposed tripeptides (M1: 1; M2: 5; M3: 3) as well as a hypothetical aaa-cys-cys tripep-tide in a reaction using aaa and cys only. Production levels vary strongly for the novel tripeptides, from 0.02 % up to 13.8 % tripeptide production rela-tive to LLD-aaa-cys-val production levels. Significantly abundant tripeptides Figure 3 — Substrate dependence of the rate of ACV synthesis.

Three reaction series were conducted using 0.025 (- - -), 0.05 (-), 0.25 (- - -), 1 (-), 2 (- - -) and 5 mM (-) of L-α-aaa (A), L-cys (B) and L-val (C) and analyzed by LC/MS to quantify the amounts of LLD-aaa-cys-val for Michaelis menten kinetics (D)

■

L-α-aaa;

▲

L-val,

●

L-cys.

0 10 20 30 40 50 60 70 0 50 100 150 200 250 0 10 20 30 40 50 60 70 0 50 100 150 200 250 0 10 20 30 40 50 60 70 0 50 100 150 200 250 LLD -A CV ( uM )

L-α-Aminoadipic acid L-Valine

L-Cysteine

time (min) time (min)

time (min) LLD -A CV ( uM ) LLD -A CV ( uM ) 0.0 0.2 0.4 0.6 0.8 1.0 1.2 0 1 2 3 4 5 Kc at (mi n-1) Velocity ACVS Substrate (mM) A B C D

(13)

100

4

Results

(No. 6; 13; 15 and 17) were further characterized for β-lactam formation through the action of purified IPNS enzyme, though none of the predicted β-lactam structures were detected, either due to product instability, sub- detection limit levels or incompatibility with the IPNS specificity.

Single particle analysis and structural evaluation of the ACVS

In order to achieve a more in depth understanding of the functional dynamics and structures of the ACVS, we further optimized the purification conditions of ACVS for a subsequent electron microscopic characterization. Initially, we

NH OH O O NH2 O S H NH CH3 C H3 OH O 1 2 OH O NH2 NH OHS NH O O H CH3 O CH3 3 O NH OHS O NH O H CH3 O CH3 4 NH OH O NH2 OHS NH O O H CH3 O CH3 5 6 7 8 NH OH O O NH2 NH CH3 O CH3 C H3 OH O OH NH O OHS NH O O H CH3 O CH3 NH O H O O NH2 NH O CH3 C H3 O H O NH OH O O NH2 NH O SH CH3 C H3 OH O 9 10 11 12 NH OH O O NH2 O H NH O CH3 C H3 OH O NH OH O O NH2 NH CH3 O O H CH3 C H3 OH O NH OH O O NH2 N H C H3 O CH3 SH CH3 CH3 OH O NH OH O O NH2 NH O S C H3 CH3 C H3 OH O 13 14 15 16 NH O H O O NH2 O NH C H3 CH3 CH3 C H3 OH O NH OH O O NH2 N H O CH3 CH3 CH3 CH3 O H O NH2 O O OH NH O SH NH O H O CH3 NH2 O O OH NH O SH NH O H O CH3 C H3 17 18 19 20 NH2 O O OH NH O SH NH O H O CH3 CH3 NH2 O O OH NH O SH NH O H C H3 O OH NH2 O O OH NH O SH NH O H O S C H3 NH2 O O OH NH O SH NH O H C H3 O CH3 SH 21 22 23* NH2 O O OH NH O SH NH OH O NH2 O O OH NH O SH NH O H CH3 O NH2 O O OH NH O S H NH CH3 C H3 OH O

Analogue Tripep�de # Mi rel prod (+-err)

ACVS - LLD-ACV 1 363.146 100 ± 10.3

L-Aspar�c acid Asp-CV 2 335.115 0 Phenoxyace�c acid POA-CV 3 354.125 0 L-Glutamine Glu-CV 4 349.131 0.02 ± 0.009 Adipic acid AA-CV 5 348.136 0 L-Alanine A-Ala-V 6 331.174 0.98 ± 0.02 L-Glycine A-Gly-V 7 317.159 0 DL-Homocysteine A-Hcys-V 8 377.162 0 L-Serine A-Ser-V 9 347.169 0 L-Threonine A-Thr-V 10 361.185 0.02 ± 0.003 L-Penicillamine A-Pen-V 11 391.178 0.05 ± 0.005 L-Methionine A-Met-V 12 391.178 0.07 ± 0.01 L-Leucine A-Leu-V 13 373.221 1.58 ± 0.21 L-Isoleucine A-IsoL-V 14 373.221 0 L-Norvaline AC-NorVal 15 363.146 13.8 ± 0.59 L-Leucine AC-Leu 16 377.162 0.54 ± 0.01 L-Isoleucine AC-IsoL 17 377.162 1.21 ± 0.04 L-Threonine AC-Thr 18 365.126 0 L-Methionine AC-Met 19 395.118 0 L-Penicillamine AC-Pen 20 395.118 0 L-Glycine AC-Gly 21 321.099 0 L-Alanine AC-Ala 22 335.115 0 L-Cysteine AC-Cys* 23 367.087 13.6 ± 1.93 M1 M2 M3

Figure 4 — Substrate promiscuity of the N. lactamdurans ACVS and structures of produced tripeptides.

Three sets of reactions were analyzed varying the amino acid on one position within the tripeptide. Structurally analogue substrates were added to a concentration of 5 mM, replac-ing either L-α-aaa (M1), L-cys (M2) or L-val (M3) (table left). Reactions were evaluated usreplac-ing LC/MS and peaks of interest were assessed according to accurate monoisotopic mass (Mi)

(14)

4

analyzed the ACVS obtained from the two-step purification using dynamic light scattering (DLS). The particle radius range of 8.4–15.7 nm is suggesting protein heterogeneity in the sample. However, a substantial amount of pro-tein aggregation occurred, as well as different conformations were present in the sample, as shown in the particle size distribution (Figure 5 A). To erad-icate the observed protein aggregation and to obtain a more homogeneous particle conformation, different additives at 1 mM concentration were eval-uated and tested (Figure 5 B–E). There was no significant change upon the addition of MgCl2 (D), but significant improvement in particle homogeneity was obtained after addition of AMP (B), aaa, cys, val (C) and a combination of these two sets of molecules (E). To further stabilize the sample and prohibit

NH OH O O NH2 O S H NH CH3 C H3 OH O 1 2 OH O NH2 NH OHS NH O O H CH3 O CH3 3 O NH OHS O NH O H CH3 O CH3 4 NH OH O NH2 OHS NH O O H CH3 O CH3 5 6 7 8 NH OH O O NH2 NH CH3 O CH3 C H3 OH O OH NH O OHS NH O O H CH3 O CH3 NH O H O O NH2 NH O CH3 C H3 O H O NH OH O O NH2 NH O SH CH3 C H3 OH O 9 10 11 12 NH OH O O NH2 O H NH O CH3 C H3 OH O NH OH O O NH2 NH CH3 O O H CH3 C H3 OH O NH OH O O NH2 N H C H3 O CH3 SH CH3 CH3 OH O NH OH O O NH2 NH O S C H3 CH3 C H3 OH O 13 14 15 16 NH O H O O NH2 O NH C H3 CH3 CH3 C H3 OH O NH OH O O NH2 N H O CH3 CH3 CH3 CH3 O H O NH2 O O OH NH O SH NH O H O CH3 NH2 O O OH NH O SH NH O H O CH3 C H3 17 18 19 20 NH2 O O OH NH O SH NH O H O CH3 CH3 NH2 O O OH NH O SH NH O H C H3 O OH NH2 O O OH NH O SH NH O H O S C H3 NH2 O O OH NH O SH NH O H C H3 O CH3 SH 21 22 23* NH2 O O OH NH O SH NH OH O NH2 O O OH NH O SH NH O H CH3 O NH2 O O OH NH O S H NH CH3 C H3 OH O

Analogue Tripep�de # Mi rel prod (+-err)

ACVS - LLD-ACV 1 363.146 100 ± 10.3

L-Aspar�c acid Asp-CV 2 335.115 0 Phenoxyace�c acid POA-CV 3 354.125 0 L-Glutamine Glu-CV 4 349.131 0.02 ± 0.009 Adipic acid AA-CV 5 348.136 0 L-Alanine A-Ala-V 6 331.174 0.98 ± 0.02 L-Glycine A-Gly-V 7 317.159 0 DL-Homocysteine A-Hcys-V 8 377.162 0 L-Serine A-Ser-V 9 347.169 0 L-Threonine A-Thr-V 10 361.185 0.02 ± 0.003 L-Penicillamine A-Pen-V 11 391.178 0.05 ± 0.005 L-Methionine A-Met-V 12 391.178 0.07 ± 0.01 L-Leucine A-Leu-V 13 373.221 1.58 ± 0.21 L-Isoleucine A-IsoL-V 14 373.221 0 L-Norvaline AC-NorVal 15 363.146 13.8 ± 0.59 L-Leucine AC-Leu 16 377.162 0.54 ± 0.01 L-Isoleucine AC-IsoL 17 377.162 1.21 ± 0.04 L-Threonine AC-Thr 18 365.126 0 L-Methionine AC-Met 19 395.118 0 L-Penicillamine AC-Pen 20 395.118 0 L-Glycine AC-Gly 21 321.099 0 L-Alanine AC-Ala 22 335.115 0 L-Cysteine AC-Cys* 23 367.087 13.6 ± 1.93 M1 M2 M3

and the resulting levels were set relative to the production of LLD-aaa-cys-val (= 100), as-suming similar ionization. The predicted structures of the novel tripeptides and their corre-sponding reactions are numbered 1–23. Italic numbers indicate production of the respective tripeptide. * = hypothetical aaa-cys-cys compound derived from L-α-aaa and L-cys only (23). Values derived from two biological and technical replicates ± standard deviation.

(15)

102

4

(16)

4

particle damage we adapted the conditions of (E) to all the buffers used in the two-step purification process, resulting in one predominant conforma-tion (F). The latter sample was subsequently subjected to buffer exchange followed by negative staining and electron microscopy (Figure 6). In addition to intact particles aggregated and partly dissociated particles were visible. It appeared that the previously compact particles enter a more disordered state upon longer incubation deeming it less suitable for detailed electron microscopic analysis (Supplementary figure 3).

Achieving conformational unity assisted by gradient fixation

To overcome issues linked to the intrinsic flexibility and instability of the ACVS, we adapted the GraFix method [23] for the stabilization of ACVS particles. Therefore, a 10–30 % glycerol gradient with and without 0.2 % glutaraldehyde (GA) was prepared, which was subsequently loaded with 100–1500 pmol of the ACVS using a one-step purified sample. The sam-ples retrieved after fractionation contained 0.15–0.40 mg/ml protein (Sup-plementary figure 2A). Although no distinct protein distribution profiles emerged after substrate addition, and the subsequent native gel analysis also showed no significant difference in the state of the ACVS, the previ-ously observed unfolding vanished (Supplementary figure 2B). Samples were subsequently subjected to negative stain analysis, showing an overall ho-mogenous picture without pronounced differences between the 0 % GA and the GA 0.2 % samples (Not shown). Negative stain samples allowed for the picking of a series of particles and their subsequent classification according Figure 5 — Dynamic light scattering (DLS) of ACVS in the presence of different additives. ACVS protein derived from the two-step purification was supplemented with various addi-tives (A–E) or purified in the presence of addiaddi-tives (F). Additive concentrations were used as indicated, and the amino acid (aa) mix contained all three native substrates (α-aaa, L-cys and L-val, 1 mM each). Samples were measured by DLS at a laser intensity of 80 % and a temperature of 4°. Particle radius (r in nm), Polydispersity (PD in %), predicted molecular weight (MW in kD), % of total mass (mass in %) and total sample peaks were determined. The most heterogeneous samples (A & D) showed Rad 15.4–15.6 nm, PD 67.3–68.5 %, MW 2011–2106 kD, and a mass of 79.1–82.5 %. Highest homogeneity was observed in (F) show-ing rad 8.4–8.6 nm, PD 4.8–11 %, MW 495–521 kD, and mass of 95.8–99.7 %. Results derived from three biological replicates and three technical replicates in addition to 10 scans per read and 10 reads per sample.

▽

= single particle;

▼

= aggregation.

(17)

104

4

Results

to their orientation on the grid, and resulted in the reconstruction of a 3-D model at a resolution of approx. 20 Å (Figure 7). At this limited resolution, it is not possible to assign domains or subdomains to the structure. However, the particle appears to consist of three modules, which are seemingly ar-ranged in a pseudo three-fold symmetry, resembling a helix-like structure that encompasses three channels (Figure 8). The particle projections from cryo-electron micrographs showed many unfolded or aggregated proteins and the resulting density map was of insufficient resolution to draw any sig-nificant conclusions on the 3- dimensional structure.

Figure 6 — Negative stain image of ACVS in the presence of AMP.

Picture taken at magnification of 100kx, indicated are the different particle conformations and orientations present in the sample.

▼

= single particle;

▽

= aggregation;

▼

= dissocia-tion/impurities. 1 = Top view of compact particle; 2 = side view of compact particle; 3 = un-folded particle.

(18)

4

Figure 7 — Different particle orientations of the ACVS.

The structures represent a 2-D average of automatically selected and subsequently grouped particles from negative stained samples.

Figure 8 — 2-D averaged top (A) and side (B) view of ACVS particle from two shifted angles. The corresponding angle of the 3-D reconstruction is shown on the right side of the

(19)

106

4

Discussion and c

onclusions

Discussion and conclusions

Here we report on the overexpression and purification of the Nocardia

lact-amdurans ACV synthetase for its biochemical characterization and structural

analysis. The gene encoding the ACVS of N. lactamdurans was heterolougsly expressed in E. coli HM0079 which contains a genomic copy of sfp, essential to activate the ACVS using a phosphopantheteine moiety. An efficient two step purification process was developed to obtain highly pure enzyme. Due to the size of over 400 kD of the protein, impurities could be minimized by size exclusion chromatography as the host cell does not expressing any intrin-sic proteins at sizes over 200 kD [24]. An initial enzymatic characterization was performed by following the production of the tripeptide LLD-aaa-cys-val. With respect to the three intrinsic ACVS modules, distinct differences in substrate affinities were noted with the initiating module showing the lowest affinity for its substrate, i.e., aminoadipate. Previous studies on the formation of product intermediates and partial reactions of NRPS enzymes suggest that the initial amino acid thiolation reaction is a rate limiting step in the assembly of nonribosomal peptides, necessary for the subsequent do-mains to adopt their distinct conformations for the peptide bond formation and product release [25–26]. Overall substrate affinities levels appear to be in line with other ACVS homologues, in particular of those of prokaryotic or-igin [27]. Adenylation domain specificities and especially the module specific velocities may also be determined in more detail, using a radiolabel based assay [28], which would allow for a more detailed comparison in relation to other characterized ACV synthetases.

We furthermore determined the substrate specificity of the ACVS mod-ules within the context of the full-length enzyme by assessing the produc-tion of tripeptides (Figure 4). Some ACVS homologues [11;29–31] exhibit a certain degree of tolerance towards substrates, despite considerably tight intrinsic control mechanisms that assure correct product formation. How-ever, with the N. lactamdurans ACVS little production was observed when aaa was replaced by potential substrate analogues. Only trace amounts of gln-cys-val resembling tripeptide were found. There seems to be no toler-ance for side chain variation, neither the absence of two side chain carbons, nor bulky-, or amino group lacking side chains were accepted. Furthermore, traceable amounts of three aaa-X-val tripeptides were found when investi-gating the module 2 (L-cysteine) promiscuity. In addition, two tripeptides, aaa-ala-val and aaa-leu-val were generated in substantial amounts. Due to the nature of the leucine and valine side chain though, it is questionable if any sort of functional downstream conversion into a biologically active

(20)

4

compound could be achieved. Trials using an isopenicillN synthetase in-deed showed no β-lactam ring formation (Data not shown). Finally, for the determination of the substrate specificity of the third module (L-valine), two novel products were observed at the level of 0.5–1 %, cys-leu and aaa-cys-ile and more importantly aaa-cys-norvaline at over 13 % of the native amounts of aaa-cys-val. It appears that there is a significant level of toler-ance towards the side chain length and the distribution of methyl-groups, however, substrates with hydroxy- or thio-groups are not incorporated at all. Further, we observed the production of the aaa-cys-cys tripeptide, which is produced under valine deficient conditions, to a level of over 13 %. Perhaps instead of a third module' cysteine incorporation however, there may be an iterative-double cysteine incorporation by module 2. The observed substrate specificity indicates further engineering potential for the N. lactamdurans ACVS. However, in order to conclusively determine the substrate specificity and potential, an extended tier of natural and non-natural substrates, such as D- or “clickable” amino acids [32–33], should be included in subsequent studies. Possibly by combinatorial methods, a powerful platform for custom NRP synthesis may be developed.

Importantly, until the intra-NRPS reaction dynamics, conformational tim-ing and substructure positiontim-ing have been conclusively elucidated, global engineering efforts will remain challenging. Thus, as part of the general char-acterization of the ACVS, we tried to evaluate and establish conditions for the purification as well as stabilization to ultimately retrieve a structurally homogenous protein, suitable for an electron microscopic analysis. There-fore, we initially attempted to utilize an ACVS extract obtained from the two step purification. Upon negative stain analysis of the particles after a buffer exchange reaction, a predominant conformation was observed, resembling in its crude shape a barrel or helix-like structure at a radius of about 10 nm. However, this sample is unstable and converts in time into a less ordered state or even an aggregated form. All potential co-factors as well as the three native substrates were added in a systematic manner, in order to stabilize the compact particle conformation. Prediction of size and mass by DLS re-vealed that ~20 % of the total protein is compact (Figure 5). Addition of mag-nesium Ions did not affect the particle size distribution, but the addition of amino acids, AMP and the combination thereof increases the compact par-ticle category to 40–50 % of the total mass. When the beneficial co-factors are present throughout the two step purification procedure, a sample with over 90 % of homogenous compact particles is obtained. Using negative stain based data collection, a 3-D reconstruction of the ACVS structure at approx-imately 20 Å resolution was generated. Trials on improving the resolution by

(21)

108

4

Ackno

wledg

ements

cryo electron microscopy failed as samples slowly converted into disordered conformations. To solve this problem, a gradient fixation or Grafix method [20;23] was applied in order to stabilize the compact conformation of the particles. Although stabilization could not be directly linked to this fixation, a generic stabilizing effect was observed due to the increased glycerol con-centration in the sample. This stabilizing effect appeared to rely on trace amounts of glycerol that remained after the buffer exchange step, but inter-fered with cryo EM. Also, the ACVS enzyme appears to bind to the carbon film leaving only low amounts of proteins available in the holes for imaging. Although there are multiple ways by which the problem of an unstable or conformational heterogeneous particles can be solved, the use of NRPS sub-structures obtained by X-ray diffraction is an efficient means to inter-pret the low resolution EM structure at a molecular level [18]. Product- and substrate intermediates and analogues, respectively, have been used to lock the conformation of distinct domains. This allowed for a more precise de-termination of intra- and inter-modular interactions and the associated sur-faces. Another approach would be the adjustment of the grid preparation method, shifting it towards a more gentle technique, such as the utilization of antibody- associated sample support grids [34–35]. Either of those ap-proaches may eventually lead to the determination of a full, multi-modular structure of a NRPS, thereby eliminating a crucial bottleneck for the system-atic engineering of NRPS as well as the downstream associated, potentially promising, bioactive compounds.

Acknowledgements

The authors would like to express their gratitude towards DSM Synochem Pharmaceuticals BV, for the Nl ACVS construct, as well as for analytical and scientific support. This work was financially supported by the BE-Basic Foun-dation, an international public–private partnership.

(22)

4 References

1. Von Döhren H, Dieckmann R, Pave-la-Vrancic M. The nonribosomal code. Chem Biol. 1999;6: 273–279.

2. Marahiel MA, Stachelhaus T, Mootz HD. Modular peptide synthetases involved in nonribosomal peptide synthesis. Chem Rev. American Chemical Society; 1997;97: 2651–2674.

doi:10.1021/cr960029e

3. Suring W, Meusemann K, Blanke A, Marien J, Schol T, Agamennone V, et al. Evolutionary ecology of beta-lactam gene clusters in animals. Mol Ecol. En-gland; 2017;26: 3217–3229.

doi:10.1111/mec.14109

4. Tahlan K, Moore MA, Jensen SE. δ-(l-α- aminoadipyl)-l-cysteinyl-d- valine synthe-tase (ACVS): discovery and perspectives. J Ind Microbiol & Biotechnol. 2017;44: 517–524. doi:10.1007/s10295-016-1850-7 5. Liempt H Van, Kleinkauf H. The first en-zyme in penicillin biosynthesis. IS; 1989; 3680–3684.

6. Baldwin JE, Bird JW, Field RA, O’Cal-laghan NM, Schofield CJ. Isolation and partial characterisation of ACV synthe-tase from Cephalosporium acremonium and Streptomyces clavuligerus. The Journal of antibiotics. Japan; 1990. pp. 1055–1057.

7. Baldwin JE, Bird JW, Field RA, O’Cal-laghan NM, Schofield CJ, Willis AC. Isola-tion and partial characterisaIsola-tion of ACV synthetase from Cephalosporium acre-monium and Streptomyces clavuligerus. Evidence for the presence of phosphop-antothenate in ACV synthetase. J Anti-biot (Tokyo). Japan; 1991;44: 241–248.

8. Jensen SE, Wong A, Rollins MJ, West-lake DW. Purification and partial char-acterization of delta-(L-alpha-amino-adipyl)-L-cysteinyl-D-valine synthetase from Streptomyces clavuligerus. J Bacte-riol. United States; 1990;172: 7269–7271. 9. Theilgaard HB, Kristiansen KN, Henrik-sen CM, NielHenrik-sen J. Purification and char-acterization of delta-(L-alpha-amino-adipyl)-L-cysteinyl-D-valine synthetase from Penicillium chrysogenum. Biochem J. England; 1997;327 ( Pt 1: 185–191. 10. van Liempt H, von Dohren H, Kleinkauf

H. delta-(L-alpha-aminoadipyl)-L-cyste-inyl-D-valine synthetase from Aspergil-lus nidulans. The first enzyme in peni-cillin biosynthesis is a multifunctional peptide synthetase. J Biol Chem. United States; 1989;264: 3680–3684. 11. Coque JJ, de la Fuente JL, Liras P, Martin JF.

Overexpression of the Nocardia lactamdu-rans alpha-aminoadipyl-cysteinyl-valine synthetase in Streptomyces lividans. The purified multienzyme uses cystathionine and 6-oxopiperidine 2-carboxylate as sub-strates for synthesis of the tripeptide. Eur J Biochem. England; 1996;242: 264–270. 12. Coque JJ, Martin JF, Calzada JG, Liras P. The cephamycin biosynthetic genes pc-bAB, encoding a large multidomain pep-tide synthetase, and pcbC of Nocardia lactamdurans are clustered together in an organization different from the same genes in Acremonium chrysogenum and Penicillium chrysogenum. Mol Microbiol. England; 1991;5: 1125–1133.

13. Keating TA, Walsh CT. Initiation, elon-gation, and termination strategies in

(23)

110

4

Ref er enc es

polyketide and polypeptide antibiotic biosynthesis. Curr Opin Chem Biol. 1999;3: 598–606.

doi:10.1016/S1367-5931(99)00015-0 14. Stein T, Vater J, Kruft V, Otto A,

Wit-tmann-Liebold B, Franke P, et al. The multiple carrier model of nonribosomal peptide biosynthesis at modular mul-tienzymatic templates. J Biol Chem. United States; 1996;271: 15428–15435. 15. Payne JAE, Schoppet M, Hansen MH,

Cryle MJ. Diversity of nature’s assembly lines — recent discoveries in non-ribo-somal peptide synthesis. Mol Biosyst. The Royal Society of Chemistry; 2017;

doi:10.1039/C6MB00675B

16. Drake EJ, Miller BR, Shi C, Tarrasch JT, Sundlov JA, Leigh Allen C, et al. Struc-tures of two distinct conformations of holo-non-ribosomal peptide synthetases. Nature; 2016;529: 235–238.

http://dx.doi.org/10.1038/nature16163 17. Marahiel MA. A structural model for

multimodular NRPS assembly lines. Nat Prod Rep. England; 2016;33: 136–140. doi:10.1039/c5np00082c

18. Tarry MJ, Haque AS, Bui KH, Schmeing TM. X-ray crystallography and electron microscopy of cross- and multi- module nonribosomal peptide synthetase pro-teins reveal a flexible architecture. Struc-ture. United States; 2017;25: 783–793.e4. doi:10.1016/j.str.2017.03.014

19. Gruenewald S, Mootz HD, Stehmeier P, Stachelhaus T. In vivo production of ar-tificial nonribosomal peptide products in the heterologous host Escherichia coli. Appl Environ Microbiol. 2004;70: 3282–3291.

doi:10.1128/AEM.70.6.3282

20. Kastner B, Fischer N, Golas MM, Sander B, Dube P, Boehringer D, et al. GraFix: sample preparation for single-particle electron cryomicroscopy. Nat Methods. United States; 2008;5: 53–55.

doi:10.1038/nmeth1139

21. Scheres SHW. Semi-automated selec-tion of cryo-EM particles in RELION-1.3. Journal of Structural Biology. 2015. pp. 114–122. doi:10.1016/j.jsb.2014.11.010 22. Rohou A, Grigorieff N. CTFFIND4: Fast

and accurate defocus estimation from electron micrographs. bioRxiv. 2015; http://biorxiv.org/content/early/ 2015/ 08/13/020917.abstract

23. Stark H. GraFix: stabilization of frag-ile macromolecular complexes for sin-gle particle cryo-EM. Methods Enzy-mol. United States; 2010;481: 109–126. doi:10.1016/S0076-6879(10)81005-5 24. Kaur J, Kumar A, Kaur J. Strategies

for optimization of heterologous pro-tein expression in E. coli: Roadblocks and reinforcements. Int J Biol Macro-mol. Netherlands; 2018;106: 803–822. doi:10.1016/j.ijbiomac.2017.08.080 25. Walsh CT. Insights into the chemical

logic and enzymatic machinery of NRPS assembly lines. Nat Prod Rep. The Royal Society of Chemistry; 2016;33: 127–135. doi:10.1039/C5NP00035A

26. Sun X, Li H, Alfermann J, Mootz HD, Yang H. Kinetics profiling of gramicidin S syn-thetase A, a member of nonribosomal peptide synthetases. Biochemistry. 2014;53: 7983–9. doi:10.1021/bi501156m 27. Schwecke T, Aharonowitz Y, Palissa

H, von Dohren H, Kleinkauf H, van Liempt H. Enzymatic characterisa-tion of the multifunccharacterisa-tional enzyme

(24)

4

delta-(L-alpha-aminoadipyl)-L-cyste-inyl-D-valine synthetase from Strepto-myces clavuligerus. Eur J Biochem. En-gland; 1992;205: 687–694.

28. Keller U, Kleinkauf H, Zocher R. 4-Meth-yl-3-hydroxyanthranilic acid activating enzyme from actinomycin-producing Streptomyces chrysomallus. Biochemis-try. United States; 1984;23: 1479–1484. 29. Zhang J, Wolfe S, Demain AL. Bio-chemical studies on the activity of delta-(L-alpha-aminoadipyl)-L-cyste-inyl-D-valine synthetase from Strepto-myces clavuligerus. Biochem J. England; 1992;283 ( Pt 3: 691–698.

30. Etchegaray A, Dieckmann R, Kennedy J, Turner G, von Döhren H. ACV synthetase: expression of amino acid activating do-mains of the Penicillium chrysogenum enzyme in Aspergillus nidulans. Bio-chem Biophys Res Commun. 1997;237: 166–169.

31. Baldwin JE, Shiau CY, Byford MF, Schofield CJ. Substrate specificity of L-delta-(alpha-aminoadipoyl)-L-cys-teinyl-D-valine synthetase from Cepha-losporium acremonium: demonstration of the structure of several unnatural tripeptide products. Biochem J. 1994;301, Pt 2: 367–72.

http://www.pubmedcentral.nih.gov/ articlerender.fcgi?artid=1137089&tool= pmcentrez&rendertype=abstract 32. Zhu X, Zhang W. Tagging polyketides/

non-ribosomal peptides with a clickable functionality and applications. Front Chem. Frontiers Media S.A.; 2015;3: 11. doi:10.3389/fchem.2015.00011 33. Park M, Wetzler M, Jardetzky TS,

Bar-ron AE. A readily applicable strategy to

convert peptides to peptoid-based ther-apeutics. Chen P, editor. PLoS One. San Francisco, USA: Public Library of Sci-ence; 2013;8: e58874.

doi:10.1371/journal.pone.0058874 34. Yu G, Li K, Jiang W. Antibody-based

af-finity cryo-EM grid. Methods. United States; 2016;100: 16–24.

doi:10.1016/j.ymeth.2016.01.010 35. Yu G, Vago F, Zhang D, Snyder JE, Yan

R, Zhang C, et al. Single-step antibody- based affinity cryo-electron microscopy for imaging and structural analysis of macromolecular assemblies. J Struct Biol. United States; 2014;187: 1–9. doi:10.1016/j.jsb.2014.04.006 36. Liras P, Demain AL. Chapter 16

Enzy-mology of beta-lactam compounds with cephem structure produced by actino-mycete. 2009. pp. 401–429.

(25)

112

4

Supplementary material

Supplementary figure 1 — Overview of ACVS reaction optimization.

The activity of ACVS in product formation was assessed under different reaction conditions.

(A) Crowding effect of BSA: black = 1 mg/ml ACVS;

■

(

■

) = 0.75 mg/ml ACVS (+0.25 mg/ml

BSA);

▲

(

▲

) = 0.5 mg/ml ACVS (+0.5 mg/ml BSA);

●

(

●

) = 0.25 mg/ml ACVS (+0.75 mg/ml BSA);

◆

(

◆

) = 0.1 mg/ml ACVS (+0.9 mg/ml BSA). (B) Effects of different agents and condi-tions to terminate the reaccondi-tions: black = 10 mM DTT; — ·· — ·· — = 75 °C; 10 min; — - — - — = 95 °C; 10 min; - · - · - = 55 °C; 10 min; --- = HCl 0.1 M; ··· = NaOH 0.1 M; grey = EDTA

0.1 M. (C) Effect of AMP. Reactions were initiated upon addition of ATP and/ or AMP. ATP + AMP mM -

●

=5/0;

■

= 4/1;

◆

= 3/2

▲

= 2/3. (D) Effect of glycerol on the ACVS activity after storage for 30 days at −80 °C: glycerol concentrations: 15 % — · · —; 25 % - - - ; 35 % ···; 50 %---. 0 20 40 60 80 100 0 50 100 150 200 250 LL D-ACV (% ) 0 20 40 60 80 100 0 200 400 600 800 1000 0 20 40 60 80 100 0 50 100 150 200 250 LL D-ACV (% ) 0 20 40 60 80 100 0 200 400 600 800 Glycerol storage AMP:ATP LL D -A CV (% ) LL D -A CV (% ) C D

�me (min) �me (min)

(26)

4

126

Supplementary figure 2 — ACVS Gradient Fixation in the presence of substrate additions (A) and fixation with glutaraldehyde (B).

6 gradients (10–30 % glycerol) were run, loading 1500 pmol of ACVS derived from a one-step purification. Collected fractions were measured at A280, represented in (A). black = ACVS only; grey = +AMP 1 mM; black dashed = +amino acids 1 mM; grey dashed = +MgCl2 1 mM; black dotted = All additives; grey dotted = All additives in buffers. Glycerol concen-tration of the corresponding fractions is indicated with the blue gradient bar. ACVS protein was also subjected to glycerol gradients with and without 0.2 % glutaraldehyde (B). Frac-tions displayed (I–VI) represent peak ACVS particle concentration samples. Every condition was evaluated from at least 2 biological replicates.

▽

= single particle;

▼

= Aggregation;

(27)

114

4

Supplementary material

Supplementary figure 3 — Different 2-D averaged images taken from cryo EM prepara-tions, representing multiple top, side and intermediate angles of the ACVS.

(28)

4

Supplementary figure 4 — pBAD plasmid containing the Nocardia lactamdurans ACVS. Parts of the 14992 bp pBAD-NlACVS-6×his plasmids have been assigned as, Nl A, C and V represent modules M1, 2 and 3 of the ACVS, pBAD promotor

▬

, terminator

■

, zeocin resis-tance marker

▭

, pBR322 origin of replication

■

, araC

▬

and gateway sites

■

.

Supplementary Figure 4 – pBAD plasmid containing the Nocardia lactamdurans ACVS

Parts of the 14992bp pBAD‐NlACVS‐6xHIS plasmids have been assigned as, Nl A, C and V represent modules M1, 2 and 3 of the ACVS, pBAD promotor �, terminator , Zeocin resistance marker �, pBR322 origin of replication , araC � and gateway sites .

(29)