University of Groningen Sustainable membrane biosynthesis for synthetic minimal cells Exterkate, Marten

(1)

Sustainable membrane biosynthesis for synthetic minimal cells

Exterkate, Marten

DOI:

10.33612/diss.98704569

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:

2019

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA):

Exterkate, M. (2019). Sustainable membrane biosynthesis for synthetic minimal cells. Rijksuniversiteit

Groningen. https://doi.org/10.33612/diss.98704569

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 3

Cardiolipin Biosynthesis

by the Promiscuous ClsA in a

Synthetic Cellular Membrane

Marten Exterkate†_{, Niels W.A. de Kok}†_{, Ruben L.H. Andringa}#_{, Iris C. Verhoek}†_{, Adriaan J.} Minnaard#_{, and Arnold J.M. Driessen}†*

† _{Department of Molecular Microbiology, Groningen Biomolecular Sciences and Biotechnology Institute, University} of Groningen, Nijenborgh 7, 9747 AG Groningen, The Netherlands

# _{Department of Chemical Biology, Stratingh Institute for Chemistry, University of Groningen, Nijenborgh 7, 9747 AG}

Groningen, The Netherlands

(3)

ABSTRACT

The bottom-up construction of a synthetic cell comprises many different processes, including expansion of the membrane. In a previously developed in vitro phospholipid biosynthesis pathway the essential phospholipid species PE and PG of the Escherichia coli cytoplasmic membrane could be produced at high levels yielding sustainable membrane growth. Here the pathway was extended with the production of cardiolipin by the E. coli cardiolipin synthase A (ClsA), thereby mimicking a native membrane composition. Further analysis of ClsA revealed a remarkable promiscuity of this enzyme towards a wide variety of substrates. In the reverse reaction that utilizes cardiolipin, numerous primary alcohols other than glycerol could be incorporated, which resulted in the synthesis of phospholipids with a wide variety in polar headgroups.

(4)

INTRODUCTION

The development of a synthetic minimal cell, that is built bottom-up from lifeless components, is a complex process in which many challenges are encountered 1,2_{. Although in such a cell}

only basic cellular functions will be implemented, still numerous cellular components have to be assembled that need to interact in a functional manner. One of the basic engineering challenges will be the generation of a boundary membrane consisting of phospholipids (mimics) that separates the compartment interior from the exterior environment 3_{. Besides}

functioning as the general barrier, such membranes additionally need to support a wide variety of membrane processes, (e.g. transport and energy transduction) 4,5_{. As often these processes}

require the presence of specific phospholipid species, the boundary layer of a synthetic minimal cell should have a well-defined lipid composition 3_{. The cytoplasmic membrane of}

Escherichia coli can serve as a suitable template, as it is relatively simple and well-studied

5_{. It consists mostly of the three main phospholipid species: phosphatidylethanolamine (PE,}

70-75%), phosphatidylglycerol (PG 20-25%) and cardiolipin (CL 0-10%, depending on the growth phase 6_{. As a first step toward a self-reproducing boundary layer of a synthetic cell,}

we recently reconstituted the E. coli phospholipid biosynthetic pathway, thereby realizing the self-assembly of enzymatically synthesized phospholipids into a growing membrane 7_.

In this system, simple fatty acid and glycerol 3-phosphate building blocks are converted by a cascade of 8 purified (membrane) proteins into the main E. coli lipid species PE and PG. In this system, cardiolipin was missing and although this phospholipid is not essential for the viability of E. coli, it is believed to facilitate several membrane related processes, among which activation of the respiratory complex and formation of intracellular membranes 8,9_.

In addition, cardiolipin appears to fulfil a key role in osmo-tolerance of membranes 10–12_.

Therefore, the introduction of cardiolipin will further extent the functional integration of a boundary layer in a synthetic cell.

Cardiolipin is an anionic phospholipid that is universally present in all domains of life. It consists of a glycerol headgroup, coupled to two pairs of acyl-chains via two glycerol-phosphate moieties, thus carrying a divalent negative charge. In bacteria CL is believed to be located in lipid-domains at the cell poles and division site 13–15_{. The molecule has an}

inverted conical structure, caused by the small anionic head group relative to the bulky hydrophobic acyl chains. Due to this structural feature, accumulation of CL possibly induces membrane curvature, but this process is not critical as CL is not an essential constituent of the E. coli membrane 16_{. In bacteria, CL is most commonly synthesized by transferring}

the phosphatidyl-group, originating from a PG molecule, to another PG, whereby glycerol leaves as a by-product. In E. coli, three enzymes are capable of producing CL: ClsA, ClsB (YbhO) and ClsC 12,17_{. Interestingly, the substrate specificity of ClsC differs from the other}

two enzymes, as ClsC utilizes the phosphatidyl group of a PE instead of another PG as donor to PG, which produces an ethanolamine leaving group 12_{. Moreover, ClsB does not}

3

(5)

only produce cardiolipin, but can also synthesizes PG, by exchanging the headgroup of PE for glycerol 18_{. Hence, depending on the substrates, an enzymatic reaction of ClsB can}

result in the production of either ethanolamine or glycerol as possible leaving group. As in E. coli, ClsA is responsible for the synthesis of the majority of CL 16,19_{, this enzyme is the}

most promising candidate for cardiolipin synthesis in a synthetic cell.

Here we report on the biosynthesis of CL and related phospholipid species by ClsA as part of an in vitro reconstituted phospholipid biosynthesis pathway to generate native-like E. coli membranes with substantial levels of CL. Furthermore, the data shows that ClsA exhibits a remarkable promiscuity that can be used to incorporate a multitude of polar headgroups into phospholipids.

RESULTS

MEMBRANE ORGANIZATION AND EXPRESSION OF ClsA

Cardiolipin synthase A (ClsA) is an integral membrane protein. Although its structure has not been determined, a membrane topology prediction based on the average hydropathy-profile of the amino acid sequences of a family of bacterial ClsA proteins reveals the presence of two hydrophobic regions at the N-terminus (corresponding to amino acids 10-29 and 36-58 in ClsA). These most likely represent transmembrane segments that are linked C-terminally to a large globular domain (Fig.1A), which is believed to associate with the membrane. This region contains two HXK motifs, which are universally present in cardiolipin synthesizing enzymes, and a common characteristic for enzymes of the phospholipase D superfamily 20_{. ClsA has a predicted mass of 54 kDa, but when a crude}

extract containing overexpressed ClsA is analyzed on SDS-PAGE, it migrates as a 46 kDa protein 17,21_{. Remarkably, in vitro transcribed and translated ClsA shows the expected mass}

of 54 kDa, suggesting post-translational processing of ClsA 22_{. Further assessment revealed}

that a N-terminal truncated (Δ2–60) ClsA variant remained membrane associated and active in vitro, whereas a C-terminal truncated ClsA is no longer functional 23_{. Therefore,}

His-tags were introduced at the N-terminus yielding His ClsA, and the aforementioned truncate His-ClsA(Δ2-60) (Fig.1B).

The genes were cloned into a pET-based overexpression vector, transformed into E. coli, and overexpressed for purification. ClsA and ClsA(Δ2-60) were found to be associated with the membrane, and could be extracted by membrane solubilization with the detergent n-dodecyl-β-d-maltoside (2%, w/v). Subsequently, Ni-NTA agarose affinity chromatography was performed, and the elution fractions were analyzed with SDS-PAGE, resulting in purified ClsA(Δ2-60) whereas His-ClsA was enriched (Fig.1C). Noteworthy, both proteins exhibit a similar mass on SDS-PAGE of about 46 kDa, suggesting truncation of the wild-type ClsA. Indeed, western blotting with a His-tag antibody confirmed the presence of the

(6)

N-terminal His-tag for ClsA(Δ2-60) while ClsA was not stained (Fig. S1A). To elucidate the processing site of ClsA, a proteomics analysis was performed using LC-MS on in-gel tryptic digests of both proteins (Fig.S1B). Peptides covering the complete amino acid sequence of ClsA(Δ2-60) could be identified, whereas no peptides were found corresponding to the N-terminal region of ClsA, which is consistent with the notion that N-terminal cleavage of ClsA occurred. To examine the functionality of both proteins, in vitro activity assays were performed in the presence of the detergent solubilized substrate PG. Whereas the truncate ClsA(Δ2-60) was largely inactive, the in vivo processed ClsA showed a high cardiolipin biosynthesis activity (Fig.1D). Control samples from a strain not overexpressing ClsA were, as expected, inactive. Hence, the enriched ClsA fraction was further used in this study. Noteworthy, introduction of a His-tag at the C-terminus of ClsA (ClsA-His) also yielded a N-terminal truncated protein, but after purification, ClsA-His was completely inactive (data not shown).

Amino acid residue

Hy drophobicit y E. coli Bacterial family ClsA His His ClsA ClsA(Δ2-60)

A

B

C

40 55 70 100 130 35 25 EV ClsA (Δ2-60)ClsA

D

Control Empty

vector ClsA ClsA(Δ2-60) 0.0 0.2 0.4 0.6 0.8 1.0 1.2 Lipid species (normlaiz ed ion count) PG CL PA MW (kDa) A C D B

Figure 1. ClsA (Ec_ClsA) and ClsA(Δ2-60). (a) Hydropathy profile alignment of E. coli ClsA (red line)

with the averaged hydropathy profile of their bacterial protein family (black line). (b) Schematic representation of ClsA and the designed mutant ClsA(Δ2-60). (c) Coomassie stained SDS-PAGE gel of purified empty vector (EV) control, ClsA and ClsA(Δ2-60) by Ni-NTA chromatography. (d) In vitro activity of purified ClsA and ClsA(Δ2-60). Lipid species were analyzed by LC-MS, normalized for the internal standard and plotted on the y-axis.

3

(7)

CARDIOLIPIN SYNTHASE ACTIVITY REPRESENTS A GLYCEROL-DEPENDENT DYNAMIC EQUILIBRIUM BETWEEN PG AND CL

In the ClsA mediated synthesis of cardiolipin, LC-MS analysis shows that about 80-85% of the PG is utilized, of which the majority is converted into cardiolipin (Fig.2A). In addition, some phosphatidic acid (PA) is produced which could arise from an unsuccessful transfer of the phosphatidyl-group, thereby directly hydrolyzing the phosphate-group. To promote formation of CL and prevent formation of PA, 100 mM of glycerol was included in the reaction. Although the production of PA was decreased slightly, the presence of glycerol resulted in higher concentrations of PG (Fig.2A), which could be caused by glycerol-mediated inhibition of ClsA. Alternatively, glycerol may stimulate the formation of PG in the reverse transesterification reaction upon hydrolysis of CL. Therefore, also the reverse reaction starting with cardiolipin was performed, in which ClsA converted CL into predominantly PG in the presence of 100 mM glycerol (Fig.2B), whereas in the absence of glycerol mostly PA was formed. This demonstrates that ClsA not only synthesizes CL from two PG molecules, but also can utilize CL to form PG and PA in a glycerol-dependent manner. Noteworthy, in the absence of glycerol, only a small amount of PG is observed, which is not proportional with the amount of PA. This suggests that the released PG originating from the hydrolysis of CL is most likely directly re-utilized for production of CL, eventually leading to the accumulation of PA that cannot be further converted. Indeed, when palmitoleoyl-oleoyl phosphatidic acid (POPA; C16:0-C18:1) was added to a reaction containing di-oleoyl phosphatidylglycerol (DOPG; C18:1-C18:1), the ClsA-mediated reaction resulted in the formation of only di-oleoyl-di-oleoyl cardiolipin (DODOCl) and DOPA, whereas no palmitoleoyl-oleoyl phosphatidylglycerol (POPG) or palmitoleoyl-oleoyl-palmitoleoyl-oleoyl cardiolipin (POPOCL) could be observed. This demonstrates that PA cannot be utilized by ClsA (Fig.S2).

The ability to distinguish a single lipid species, such as PG, based on variations in the acyl-chain composition was further exploited to examine the Cl biosynthetic equilibrium reaction and to address the transesterification step. Incubation of ClsA with an equimolar ratio of POPG and DODOCL, in the presence of glycerol, resulted in the expected production of POPOCL, DOPG and DOPA, but also in the production of POPA (Fig.2C), thereby demonstrating ClsA mediated conversions occur in both directions (Fig.2D). Moreover, a cardiolipin species with the mixed DOPO acyl-chain configuration could be identified as well. (Fig.2C). This molecule is the synthesis product of a POPG combined with a DOPG, which originated from the hydrolytic or transesterification reaction of DODOCL, thereby emphasizing the dynamic character of this equilibrium.

(8)

2 PG CL PG+PA Glycerol Glycerol H2O

A

B

C

POPG; DOPG (normaliz ed ion count) Control ClsA 0.0 0.5 1.0 0.0 0.05 0.1 POPG POPOCL POPA DOPG DODOCL DOPA PODOCL

Other lipid species

(normaliz

ed ion count)

Lipid species

(normaliz

ed ion count)

Control ClsA ClsA +

glycerol 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 PG CL PA Lipid species (normaliz ed ion count)

Control ClsA ClsA +

glycerol 0.0 0.2 0.4 0.6 0.8 PG CL PA

D

A

B

C

Other lipid species

(normaliz

ed ion count)

Lipid species

(normaliz

ed ion count)

Control ClsA ClsA +

glycerol 0.0 0.2 0.4 0.6 0.8 PG CL PA

D

A

B

C

Other lipid species

(normaliz

ed ion count)

Lipid species

(normaliz

ed ion count)

Control ClsA ClsA +

glycerol 0.0 0.2 0.4 0.6 0.8 PG CL PA

D

A

B

C

Other lipid species

(normaliz

ed ion count)

Lipid species

(normaliz

ed ion count)

Control ClsA ClsA +

glycerol 0.0 0.2 0.4 0.6 0.8 PG CL PA

D

C D B A

Figure 2. ClsA activity in vitro in the presence or absence of glycerol starting with the substrate(s) (a) PG, (b) CL and (c) POPG together with CL. Lipid species were analyzed by LC-MS, normalized for

the internal standard and plotted on the y-axis. (d) Schematic representation of the ClsA mediated glycerol-dependent dynamic equilibrium.

ENZYMATIC RECYCLING OF CL-DERIVED PA IN THE PHOSPHOLIPID BIOSYNTHETIC PATHWAY

The production of PA during ClsA mediated cardiolipin synthesis is an undesired phenomenon, as it is a dead-end product in the reversed reaction that will accumulate. However, PA is a common intermediate in phospholipid biosynthesis, and in principle would be recycled into the pathway by the enzyme CdsA that utilizes PA and CTP to yield CDP-DAG, a central precursor in phospholipid synthesis. CDP-DAG can then be converted into PG through the intermediate PGP via the enzymes PgsA and PgpA, which requires one molecule of glycerol 3-phosphate. This concept was tested by the co-reconstitution of ClsA with CdsA, PgsA and PgpA. In this reaction, the PA was completely converted into PG and cardiolipin (Fig.3A), whereas in the absence of ClsA, PA was completely converted into PG. Since in both reactions, no PA could be detected, complete recycling of the PA that originated from the hydrolysis of CL must have occurred.

The aforementioned reaction was expanded with the other enzymes of the in vitro phospholipid synthesis pathway yielding PA starting from simple building blocks 7_{. Herein,}

oleic acid is converted by the enzyme FadD into acyl-CoA. This acyl-chain donor is utilized

3

(9)

by the enzymes PlsB and PlsC, which attach an acyl-chain onto respectively the 1- and the 2-position of glycerol-3-phosphate (G3P), resulting in the formation of PA. As shown before, combining all the enzymes used in the aforementioned pathway in vitro, resulted in almost complete conversion of fatty acid into PG (Fig.3B). However, some cardiolipin production can be observed as well, possibly due to co-purification. In the presence of only CdsA, PgsA and PgpA, no cardiolipin was observed, indicating that ClsA would co-purify with either FadD, PlsB or PlsC. Nevertheless, by addition of ClsA, a significantly larger fraction of PG is converted into cardiolipin, thereby confirming successful incorporation of ClsA into the in vitro phospholipid biosynthesis pathway. Furthermore, PG-CL ratios are similar starting from the building block PA or oleic acid, indicating full compatibility of ClsA with the pathway (Fig.3).

A

B

Control PG synthesis CL synthesis

0 1 2 3 4 5 6 7 Lipid species (normaliz ed ion count) Oleic acid PG CL

0.0 0.2 0.4 0.6 0.8 1.0 Lipid species (normaliz ed ion count) PA PG CL

A

B

0 1 2 3 4 5 6 7 Lipid species (normaliz ed ion count) Oleic acid PG CL

0.0 0.2 0.4 0.6 0.8 1.0 Lipid species (normaliz ed ion count) PA PG CL A B

Figure 3. Introduction of Cardiolipin synthesis in the in vitro phospholipid biosynthesis pathway starting

from either (a) the intermediate PA, or (b) oleic acid. Lipid species were analyzed by LC-MS, normalized for the internal standard and plotted.

ClsA EXHIBITS A HIGH PROMISCUITY TOWARD PRIMARY ALCOHOLS

In the reverse CL hydrolysis reaction catalyzed by ClsA, transesterification results in the incorporation of a glycerol headgroup onto PA yielding PG. To examine the specificity of ClsA towards different alcohols, molecules structurally related to glycerol were tested as a substrate in the CL hydrolysis reaction (Fig.4A). Conversion of CL in the presence of 1-propanol resulted in the production of PG and phosphatidyl-1-propanol (p-1-prOH), whereas 2-propanol appeared a poor substrate in this reaction. This indicates that a terminal OH-group is essential. Likewise, 1,2-propanediol and 1,3-propanediol could function as substrates for synthesis of their phosphatidyl-analogue phosphatidyl-propanediol. However, in the case of 1,3-propanediol the cardiolipin analogue di-phosphatidyl-1,3-propanediol (di-p-1,3-prOH), which contains a 1,3-di-phosphatidyl-1,3-propanediol headgroup instead of a glycerol, was produced as well. This demonstrates that not only phosphatidyl-glycerol (PG), but also phosphatidyl-1,3-propanediol (p-1,3-prOH) can serve as an acceptor in the ClsA

(10)

mediated phosphatidyl-transfer.

To further test the promiscuity of the enzyme, methylated (2,2-dimethyl-1,3-propanediol) and phenylated (2-phenyl-1,3-propanediol) derivatives of 1,3-propanediol were tested as substrates (Fig.4B). Although this resulted in the production of both phosphatidyl-2,2-dimethyl-1,3-propanediol (p-2,2-Me-1,3-prOH) and phosphatidyl-2-phenyl-1,3-propanediol (p-2-Phe-1,3-prOH), the latter was produced at only low quantities. Moreover, in presence of 2,2-dimethyl-1,3-propanediol, production of PG and PA was reduced, and nearly completely absent in the presence of 2-phenyl-1,3-propanediol, which indicates steric hindrance at the second carbon impairs the overall hydrolytic enzyme activity. A similar effect of these two substrates is observed during the synthesis of cardiolipin, indicating that the steric hindrance inhibits the overall activity of the enzyme (fig S3A).

Like the tri-carbon alcohols, also di- and tetra-carbon alcohols could be used by ClsA to produce PG analogues, which is illustrated by the ethanediol-, 1,3-butanediol- and 1,4-butanediol-dependent synthesis of phosphatidyl-1,2-ethanediol (p-1,2-etOH), phosphatidyl-1,3-butanediol (p-1,3-buOH) and phosphatidyl-1,4-butanediol (p-1,4-buOH), respectively (Fig.4A). However, only in the presence of the tetra-carbon alcohols the cardiolipin analogues di-phosphatidyl-1,3-butanediol-cardiolipin (di-p-1,3-buOH) and 1,4-butanediol-cardiolipin (di-p-1,4-buOH) could be observed. As only trace amounts of di-p-1,3-buOH (2% of di-p-1,4-buOH) could be detected, ClsA seems to function better with the substrate p-1,4-buOH, suggesting that a second primary alcohol is preferred as phosphatidyl-acceptor. The promiscuity of ClsA to a diversity of alcohols is best exemplified by its ability to incorporate mannitol 24_{, which allows for the formation of both}

phosphatidyl-mannitol (p-phosphatidyl-mannitol) and di-phosphatidyl-phosphatidyl-mannitol (di-p-phosphatidyl-mannitol) (Fig.4A). This sugar comprises 6 carbons, all attached to an alcohol moiety, illustrating that ClsA is capable of incorporating a wide variety of alcohols as polar headgroup, resulting in the synthesis of a diverse group of phospholipids.

Besides the glycerol related alcohols, some polar headgroups of lipids commonly found in nature were tested as substrates as well (Fig.4C). Serine, ethanolamine, choline and inositol were tested as alcohol donors in the CL hydrolysis reaction. Although PS, PE, PC and PI could be observed by LC-MS analysis, they are only produced in minute amounts, hence hardly any CL was consumed. Notably, in the presence of ethanolamine the common hydrolysis of CL into PG and PA was reduced, indicating an inhibiting effect of ethanolamine. A similar observation was made by using 3-amino-propanol, suggesting that also this compound impairs the ClsA hydrolytic activity (Fig.S3B). Vice versa, the synthesis of CL in the presence of either ethanolamine or 3-aminopropanol reduces CL synthesis as well, illustrating that the overall activity of ClsA is impaired (Fig.S3C).

3

(11)

RT (min) 0 5 10 15 20 25 0 2.0x106 4.0x106 6.0x106

Lipid species (Ion count)

741.54 m/z p-1-prOH 0 5 10 15 20 25 0 1x104 2x104

741.54 m/z p-2-prOH 0 5 10 15 20 25 0 5.0x106 1.0x107 1.5x107 2.0x107

757.53 m/z p-1,2-prOH 757.53 m/z 7 0 5 10 15 20 25 0 1x107 2x10

Lipid species (Ion count) 1440.02 m/z

p-1,3-prOH di-p-1,3-prOH 0 5 10 15 20 25 0 5.0x106 1.0x107 1.5x107 2.0x107

743.52 m/z p-1,2-etOH 0 2.0x106 4.0x106 6.0x106 8.0x106 0 1x104 2x104 3x104 4x104 0 5 10 15 20 25

p-1,3-buOH (Ion count) di-p-1,3-buOH (Ion count) p-1,3-buOH 1454.04 m/z 771.55 m/z di-p-1,3-buOH 0 5 10 15 20 25 0 1x107 2x107 0 2.0x105 4.0x105 6.0x105

p-1,4-buOH (Ion count) di-p-1,4-buOH (Ion count)

771.55 m/z 1454.04 m/z p-1,4-buOH di-p-1,4-buOH 0 5 10 15 20 25 0 5.0x106 1.0x107 1.5x107

863.56 m/z

1546.05 m/z

p-mannitol

di-p-mannitol

R-group

Control ClsA ClsA +

1,3-prOH glycerolClsA + 2,2-Me-ClsA + 1,3-prOH ClsA + 2-Phe-1,3-prOH 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Lipid species (normaliz ed ion count) CL PG PA p-1,3-prOH di-p-1,3-prOH p-2,2-Me-1,3-prOH p-2-Phe-1,3-prOH

A

B

C

RT (min)

Control ClsA ClsA +

choline ClsA +serine ClsA +inositol ethanol-ClsA + amine 0.0 0.1 0.2 0.3 0.4 0.5 CL PG PA PC PS PI PE Lipid species (normaliz ed ion count) B A C

Figure 4. in vitro ClsA activity in the presence of various alcohols. (a) Glycerol-like alcohols: 1-propanol,

2-propanol, 1,2-propanediol, 1,3-propanediol, 1,2-ethanediol, 1,3-butanediol, 1,4-butanediol and mannitol. (b) 1,3-propanediol and its derivatives: glycerol (1,2,3-propanetriol), 2,2-dimethyl-1,3-propanediol and 2-phenyl-1,3-2,2-dimethyl-1,3-propanediol. (c) Common polar lipid headgroup alcohols: choline, serine, inositol, ethanolamine. Lipid species were analyzed by LC-MS, if noted normalized for the internal standard and plotted.

(12)

ClsA ACTIVITY WITH THE ARCHAEAL ARCHAETIDYLGLYCEROL

In archaea, phospholipids consist of highly-methylated isoprenoid chains that are ether-linked to a glycerol-1-phosphate backbone, whereas in bacteria and eukaryotes, fatty acid chains are ester bonded to the enantiomeric backbone glycerol-3-phosphate 25_.

Thus, the bacterial and archaeal phospholipids are chiral distinct. To further examine the substrate promiscuity, ClsA (Ec_ClsA) was incubated with archaetidylglycerol (AG), the archaeal equivalent of PG. This resulted in the formation of an archaeal cardiolipin (aCL) equivalent and the byproduct archaeatidic acid (AA) (fig 5A). Subsequently, a mixture of PG and AG was applied. Although in this case ClsA did not synthesize aCL, two other distinct cardiolipin species were produced. (Fig.5B). Besides bacterial cardiolipin (bCL), an additional cardiolipin species containing both isoprenoid lipid-tails and fatty acid lipid-tails was observed, thereby forming an archaeal-bacterial hybrid cardiolipin (hCL). Interestingly, not only aCL, but also AA was absent, indicating no hydrolysis of AG, thereby suggesting that ClsA prefers PG over AG as phosphatidyldonor in CL synthesis. Nevertheless, ClsA appears highly promiscuous as it also accepts the archaeal AG as a substrate.

aCL homologs are found in some Archaea, most notably in certain halophiles and methanogens, organisms that belong to the clade of euarchaeota. To identify a possible cardiolipin synthesizing enzyme in archaea, we performed a BLAST homology search with ClsA which revealed 2 main clusters of archaeal homologs with phospholipase D like activity (Fig.S4A). One cluster consists of halophilic archaea and the other concerns methanogenic archaea. The putative cardiolipin synthase from Methanospirilum hungatei was selected for further analysis. Although this enzyme only shows 27% sequence identity with the E. coli ClsA, its average hydropathy-profile also predicts two hydrophobic regions at the N-terminus, thereby showing a high structural similarity between these two proteins. M. hungatei is a strict anaerobe, growing at mesophilic temperatures with a pH around 7. Additionally, the salt requirement of M. hungatei is more comparable to E. coli. The M. hungatei ClsA homologue (Mh_Cls) was cloned into an overexpression vector and, after expression in E. coli, purified (Fig.S4B) and tested in vitro. The Mh_Cls more efficiently converted AG into aCL as compared to ClsA, thereby confirming its functionality as an archaeal cardiolipin synthesizing enzyme (Fig.5C). Moreover, by feeding Mh_Cls with both PG and AG not only archaeal CL (aCL), but also the hybrid cardiolipin (hCL) and bacterial cardiolipin (bCL) could be synthesized, indicating a high degree of promiscuity for this archaeal enzyme as well (Fig.5D).

3

(13)

Control Ec_ClsA 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 PG AG PA AA bCL aCL hCL Control Ec_ClsA 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 Lipid species (normaliz ed ion count) AG AA aCL Control Mh_Cls 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 AG AA aCL Control Mh_Cls 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 PG AG PA AA bCL aCL hCL Lipid species (normaliz ed ion count) Lipid species (normaliz ed ion count) Lipid species (normaliz ed ion count)

A

B

C

D

D B A C

Figure 5. Activity of the bacterial Ec_ClsA in the presence of (a) AG (b) together with PG, compared to

the activity of the archaeal Mh_Cls in the presence of (c) AG (d) together with PG. Lipid species were analyzed by LC-MS, normalized for the internal standard and plotted.

DISCUSSION

Here, we report on the in vitro activity of ClsA, the primary cardiolipin synthesizing enzyme in E. coli. By reconstituting this enzyme into liposomes, we were able to study this enzyme in the absence of its native membrane and in a controlled environment. In vitro activity assays in the presence of PG and/or CL revealed there is a dynamic equilibrium between these lipid species, which is, among others, depending on the amount of glycerol. The hydrolysis of cardiolipin does not only yield PG, but also results in the synthesis of the by-product PA. As PA cannot be reutilized by ClsA, we developed a recycling system that converts PA back into PG via an enzymatic cascade consisting of the three proteins CdsA, PgsA and PgpA. This recycling system will allow for sustainable phospholipid biosynthesis, without accumulation of the by-product PA. Because of the recycling additional energy is consumed, as CTP and G3P are lost in a futile cycle. This enzymatic cascade was further extended with the enzymes FadD, PlsB and PlsC, thereby allowing the synthesis of cardiolipin from simple fatty acid and glycerol building blocks. Together with PssA and PsD 7_{, this enables a complete in vitro}

mimic of phospholipid biosynthesis in E. coli that may serve as a phospholipid-membrane expanding system in a synthetic cell.

Besides cardiolipin biosynthesis, ClsA can potentially also generate other phospholipid

(14)

species. The hydrolysis of cardiolipin is not exclusive for the glycerol-dependent production of two PG molecules, but allows for a wide variety of phosphatidyl-lipids, depending on the incorporated alcohol. This phenomena has been observed in vivo with mannitol 24_and

other alcoholic substrates 26_{. In principal, any primary alcohol can serve as a}

phosphatidyl-headgroup, provided that it is compatible with the binding pocket of ClsA. This is illustrated by introducing different side-groups at the second carbon of 1,3-propanediol. Depending on the specific side-group (size, hydrophobicity, etc.), ClsA incorporates these alcohols with different efficiency. Interestingly, the length of the carbon-chains seems irrelevant, as ClsA could incorporate ethanediol, a di-carbon, as well as the hexa-carbon mannitol. Remarkably, in the presence of extracellular butanol, E. coli was found to produce phosphatidyl-butanol in a DOPC-dependent manner 26_{. As this lipid species is not endogenously present in}_E.

coli, this observation is rather difficult to explain. Moreover, in our in vitro setup, DOPC was not needed for the production of phosphatidyl-butanol from cardiolipin in the presence of butanol.

The ability of ClsA to incorporate a wide variety of headgroups, comes with new bio-catalytic applications of this enzyme, i.e. for the synthesis of natural relevant phospholipid species. Several alcohols that commonly appear as phospholipid headgroups were tested with ClsA, which allowed for the synthesis of PE, PC, PS and PI. Although these phospholipid species could so far only be produced in miniature amounts, the sensitivity of ClsA to such a wide variety of primary alcohols, could make it an interesting candidate to generate a membrane with a diverse phospholipid-headgroup composition in synthetic cells. However, for this purpose further engineering is required.

The versatility of ClsA is not only limited to a wide variety of alcohol-headgroups, as the enzyme appears to be invariant to the phospholipid-tail, as well as the glycerol-backbone. This is demonstrated by the ability of ClsA to recognize the archaeal phosphatidylglycerol (AG), which contains isoprenoid chains that are ether-linked to a G1P backbone, and convert it into archaeal cardiolipin (aCL). Moreover, when PG and AG were supplied as a mixed substrate, a novel hybrid-cardiolipin species was produced that contains one pair of the archaeal-isoprenoid chains and one pair of the bacterial-fatty acid chains. Likewise, an archaeal cardiolipin synthase, derived from Methanospirillum hungatei (Mh_Cls), showed a similar promiscuity but was more efficient in the formation of the aCL. In this respect, other polar head group attaching enzymes from archaeal and/or bacterial sources also appear invariant to the glycerol core and hydrophobic chains 27_{. This sheds new light on the ‘lipid}

divide’ theory, which states that the lipidome is a crucial and distinct marker separating the archaeal and bacterial domains of life. The ability of certain enzymes originating from both domains of life to cross recognize substrates originating from the other domain of life, further indicates that the later steps in phospholipid biosynthesis, i.e., polar headgroup attachment and modifications are based on enzymatic reactions that are functionally equivalent while the enzymes are structurally related.

3

(15)

MATERIALS AND METHODS

BIOINFORMATIC IDENTIFICATION OF Mh_Cls

Using E. coli K12; MG1655 ClsA (Ec_ClsA: NP_415765.1) as the query sequence, BLAST homology searches to the domain of Archaea were performed. From these hits a putative cardiolipin synthase (Mh_Cls: NP_415765.1) from Methanospirilum hungatei JF-1 was selected.

BACTERIAL STRAINS AND CLONING PROCEDURES

Genomic DNA of Escherichia coli was used as a template for the amplification of genes encoding for the enzymes ClsA (Ec_ClsA) and ClsA(Δ2-60). An E. coli codon-optimized synthetic gene of M. hungatei Cls (Mh_Cls) was ordered (GeneArt, Thermo Fisher scientific). and used as a template for the amplification of Mh_Cls.

E. coli DH5α (Invitrogen) was used for cloning. Plasmids for the E. coli enzymes FadD, PlsB, PlsC, CdsA, PgsA and PgpA were as reported 7_{. All primers and plasmids used in}

the present study are listed in Tables 1 and 2. E. coli BL21 (DE3) was used as a protein overexpression host strain for FadD, PlsB, PlsC, CdsA, PgsA and PgpA and E. coli Lemo21 (DE3) protein as overexpression host strain for ClsA, ClsA(Δ2-60) and Mh_Cls. All strains were grown under aerobic conditions at 37◦_{C in LB medium supplemented with the required}

antibiotics, kanamycin (50 μg/ml), chloramphenicol (34 μg/ml) and ampicillin (50 μg/ml).

EXPRESSION AND PURIFICATION OF PHOSPHOLIPID SYNTHESIZING ENZYMES

The proteins FadD, PlsB, PlsC, CdsA, PgsA and PgpA were expressed and purified as described 7_{. ClsA, ClsA(Δ2-60) and Mh_Cls were overexpressed in}_{E. coli Lemo21 (DE3)}

strain and induced with 0.5 mM IPTG. After 2.5 hrs. of induction cytoplasmic and membrane fractions were separated as described 28_{. The total membranes fractions were resuspended}

in buffer A (50 mM Tris/HCl, pH 8.0, 100 mM KCl and 15% glycerol) after which they could be stored at -80. For further purification, 0.5 mg/ml of membranes were solubilized in 2% n-dodecyl-β-D-maltopyranoside (DDM) detergent for 1 hr. at 4◦_{C. The material was subjected}

to a centrifugation (17000xg) step for 15 min at 4◦_{C to remove insolubilized material and the}

supernatant was incubated with Ni-NTA (Ni2+_{- nitrilotriacetic acid) beads (Sigma) for 2 hrs.}

at 4◦_{C. The Ni-NTA beads were washed 10 times with 40 column volumes (CV) of buffer B (50}

mM Tris/HCl, pH 8.0, 100 mM KCl, 15% glycerol and 0.05% DDM) supplemented with 10 mM imidazole, and the proteins were eluted three times with 0.5 CV of buffer B supplemented with 300 mM imidazole. Purity of the eluted proteins were assessed on 15% SDS/PAGE stained with Coomassie Brilliant Blue and the protein concentration was determined by measuring the absorbance at 280 nm. Extinction coefficients were obtained from the ProtParam tool from ExPASy by applying the specific amino acid sequence.

(16)

Table 1. Cloning and expression vectors used in this study.

Plasmid Description Reference

pRSF-Duet-1 Expression vector (KanR_{), T7 promoter} _Novagen

pET-Duet-1 Expression vector (AmpR_{), T7 promoter} _Novagen

pET-28b Expression vector (KanR_{), T7 promoter} _Novagen

pACYC-Duet-1 Expression vector (CmR_{), T7 promoter} _Novagen

pME001 fadD gene with C-terminus His-tag from E. coli MG1655 cloned into pRSF-Duet-1 vector using the primers PrME001 and PrME002

Exterkate et al.a

pME002 plsB gene with C-terminus His-tag from E. coli MG1655 cloned into pet-28b vector using the primers PrME003 and PrME004

Exterkate et al.a

pME003 plsC gene with C-terminus His-tag from E. coli MG1655 cloned into pet-28b vector using the primers PrME005 and PrME006

Exterkate et al.a

pME004 ClsA (Ec_ClsA) gene with N-terminus His-tag from E. coli MG1655 cloned into pRSF-Duet vector using the primers PrME007 and PrME008.

This study pME005 ClsA(Δ2-60) gene with N-terminus His-tag from E. coli MG1655 cloned into

pRSF-Duet vector using the primers PrME009 and PrME008.

This study pNDK001 Mh_Cls gene with N-terminus His-tag from M. hungatei JF-1 cloned into

pRSF-Duet vector using the primers NDKo027 and NDKo028.

This study

pSJ148 cdsA gene with N-terminus His-tag from E.

E. coli MG1655 cloned into pACYC-Duet vector using the primers 103 and 106

Caforio et al.b

pAC015 pgsA gene with C-terminus His-tag from E. coli MG1655 cloned into pRSF-Duet vector using the primers 551 and 552

Caforio et al.b

pAC017 pgpA gene with C-terminus His-tag from E. coli MG1655 cloned into pET-Duet vector using the primers 562 and 563

Caforio et al.b

a 7; b 27

Table 2. Oligonucleotide primers used in this study.

Primers Primer sequence 5’ -> 3’ Restriction _site

PrME001 TACTAGGAATTCATGAAGAAGGTTTGGCTTAACCG EcoRI

PrME002 AGTCATGCGGCCGCTCAGTGGTGGTGGTGGTGGTGGGCTTTATTGTCCACTTTGC NotI

PrME003 CATCCTGCCCATGGCCGGCTGGCCACGAATTTACTAC NcoI

PrME004 CTTCATGATTCTCGAGCCCTTCGCCCTGCGTCGCAC XhoI

PrME005 CACTACGACCATGG TATATATCTTTCGTCTTATTATTACCG NcoI

PrME006 CATGTCTACTCGAGAACTTTTCCGGCGGCTTC XhoI

PrME007 CGTAGCATATGCACCACCACCACCACCACACAACCGTTTATACGTTGGTGAGTTGGTTGGCCATTCTGGG NdeI

PrME008 ACGTCCTCGAGTTACAGCAACGGACTGAAG XhoI

PrME009 CTAGTCATATGCACCACCACCACCACCACCTCCATTTAGGCAAACGCCG NdeI

NDKo027 ACAGTTCCATGGCCCATCACCATCATCACCACATCCATGATCTGATTCTGGTGATCCACAATTTTC NcoI

NDKo028 ACTTACGAGCTCTTATTATTACAGCAGCGGACTAAACAGACG SacI

103 GCGCCTCGAGTTAGTGATGGTGATGGTGGTGATGATGAAGCGTCCTGAATACCAGTAAC XhoI 106 GCCGCCATGGGCAGCCATCACCATCATCACCACAGCCTGAAGTATCGCCTGATATCTGC NcoI 551 GCGCCATATGCAATTTAATATCCCTACGTTGC NdeI 552 CGCGCTCGAGTCAGTGATGGTGATGGTGGTGATGATGCTGATCAAGCAAATCTGCACGC XhoI 562 CGCGGAATTCATGACCATTTTGCCACGCCATAAAG EcoRI 563 CGGCGCGGCCGCCTAGTGATGGTGATGGTGGTGATGATGCGACAGAATACCCAGCG NotI

3

(17)

LIPOSOMES PREPARATION

Chloroform stocks of the lipid species DOPG, Cardiolipin (CL), POPG, DOPA, DOPE, DOPC were purchased from Avanti (Avanti Biochemicals, Birmingham, AL), whereas AG (Fig. S5) was synthesized by Ruben L.H. Andringa, from the Chemical Biology Department, and will be detailed elsewhere. For liposomes with a heterogeneous lipid mixture, the desired lipid chloroform stocks were mixed together in the stated ratio. Next the lipid solution was dried under a nitrogen gas stream for multiple hrs., after which the dry lipid film was resuspended in a 300mM KPi, pH 7.0 buffer. For formation of liposomes, a sonication cycle of 15 sec on, 15 sec off was repeated (10-40x) till the solution became transparent.

IN VITRO ASSAYS FOR PHOSPHOLIPID PRODUCTION

All in vitro reactions were performed in 100 μl of assay buffer A containing a final concentration of 300 mM potassium phosphate pH 7.0, 0.1% DDM and 2 mM DTT, except for the combined activity of ClsA with the enzymes derived from the earlier described in vitro phospholipid biosynthesis pathway, which was performed in buffer B 100 mM MES, pH 7.0, 10 mM MgCl₂, 100 mM KCl, 2 mM DTT. The activity of ClsA in comparison to ClsA(Δ2-60) was assayed in buffer A using 1 μM enzyme and 1 mM of DOPG. The glycerol-dependent dynamic equilibrium of ClsA was assayed in buffer A using 1 μM enzyme and either 1 mM DOPG, 0.5 mM CL, 1 mM POPG together with 0.5 mM CL, or 1 mM DOPG together with 1 mM POPA.

For recycling of PA, buffer B was supplemented with 3 mM liposomes (DOPE,DOPC,DOPA; ratio 1:1:1), 2 μM CdsA, 4 mM CTP, 2 μM PgsA, 1 μM PgpA and 1 μM ClsA. Conversion of oleic acid into PG and CL was assayed under the same conditions, except liposomes without DOPA were used, and the following compounds were additionally added: 2 μM FadD, 50 μM CoA, 2700 μM oleic acid, 4 mM ATP, 9 mM G3P, 2 μM PlsB, 3 μM PlsC.

The promiscuity toward primary alcohols was assayed in buffer A supplemented with 1 μM ClsA or Mh_Cls, 0.5 mM CL or 1 mM DOPG, and 100 mM alcoholic substrate. The activity of ClsA (Ec_ClsA) and Mh_Cls with AG/PG was assayed in buffer A with 100 μM of lipid substrate(s) and 1 μM of enzyme.

All reactions were incubated overnight at 37◦_{C unless stated differently. Lipids were}

extracted two times with 0.3 ml of n-butanol, and evaporated under a stream of nitrogen gas and resuspended in 50 μl of methanol for LC-MS analysis.

LC–MS ANALYSIS OF LIPIDS

Samples from the in vitro reactions were analyzed using an Accela1250 high-performance liquid chromatography (HPLC) system coupled with an electrospray ionization mass spectrometry (ESI–MS) Orbitrap Exactive (Thermo Fisher Scientific). A sample of 5 μl was injected into an ACQUITY UPLC® CSH™ C18 1.7μm Column, 2.1 x 150 mm (Waters Chromatography Ireland Ltd) operating at 55◦_{C with a flow rate of 300 μl/min. Separation of}

the compounds was achieved by a changing gradient of Mobile phase A (5 mM ammonium

(18)

formate in water/acetonitrile 40:60, v/v), and mobile phase B (5 mM ammonium formate in acetonitrile/n-butanol, 10:90, v/v). The column effluent was injected directly into the Exactive ESI-MS Orbitrap operating in negative ion mode. Voltage parameters of 3 kV (spray), −75 V (capillary), −190 V (tube lens) and -46 V (Skimmer voltage) were used. Capillary temperature of 300°C, sheath gas flow 60, and auxiliary gas flow of 5 was maintained during the analysis. Spectral data constituting total ion counts were analyzed using the Thermo Scientific XCalibur processing software by applying the Genesis algorithm based automated peak area detection and integration. The total ion counts of the extracted lipid products: oleic acid (m/z 281.25 [M-H]-_{), LPA (C18:1) (m/z 435.26 [M-H]}-_{), PA/DOPA (C18:1-18:1) (m/z 699.49 [M-H]}-_),

CDP-DAG (m/z 1004.54 [M-H]-_{), PGP (m/z 853.50 [M-H]}-_{), PG/DOPG (C18:1-18:1) (m/z 773.53}

[M-H]-_{), POPA (C18:1-16:0) (m/z 673.48 [M-H]}-_{), POPG (C18:1-16:0) (m/z 747.52 [M-H]}-_{), CL/}

DODOCL/bCL (C18:1-18:1-18:1-18:1) (m/z 1456.03 [M-H]-_{), POPOCL (C16:0-18:1-16:0-18:1)}

(m/z 1403.99 [M-H]-_{), DOPOCL (C18:1-18:1-16:0-18:1) (m/z 1430.01 [M-H]}-_{), DOPE (m/z 742.54}

[M-H]-_{), p-NH2-prOH (m/z 756.55 [M-H]}-_{), DOPC (m/z 830.59 [M-H]}-_{), DOPS (m/z 786.53 [M-H]}

-), DOPI (m/z 861.55 [M-H]-_{), p-2,2-Me-1,3-prOH (m/z 786.56 [M-H]}-_{), p-2-Phe-1,3-prOH (m/z}

833.56 [M-H]-_{), p-1-prOH and p-2-prOH (m/z 741.54 [M-H]}-_{), p-1,2-prOH and p-1,3-prOH (m/z}

757.53 [M-H]-_{), di-p-1,3-prOH (m/z 1440.02 [M-H]}-_{), p-1,2-etOH (m/z 743.52 [M-H]}-_),

p-1,3-buOH and p-1,4-p-1,3-buOH (m/z 771.55 [M-H]-_{), di-p-1,3-buOH and di-p-1,4-buOH (m/z 1454.04}

[M-H]-_{), p-mannitol (m/z 863.56 [M-H]}-_{), di-p-mannitol (m/z 1546.05 [M-H]}-_{), AG (m/z 805.66}

[M-H]-_{), AA (m/z 731.63 [M-H]}-_{), aCL (m/z 1520.29 [M-H]}-_{), hCL (m/z 1488.15 [M-H]}-_{), were}

normalized for either DDM (m/z 509.3 [M-H]-_{) or didecanoyl-PG (C10:0-10:0) (m/z 553.31}

[M-H]-_{) and plotted on the y-axis as normalized ion count.}

ACKNOWLEDGEMENTS

This work was supported by the NWO Gravity programme BaSyC and the NWO Building blocks of life programme.

3

(19)

REFERENCES

1. Schwille, P.; Spatz, J.; Landfester, K.; Bodenschatz, E.; Herminghaus, S.; Sourjik, V.; Erb, T. J.; Bastiaens, P.; Lipowsky, R.; Hyman, A.; et al. MaxSynBio: Avenues Towards Creating Cells from the Bottom Up. Angewandte Chemie (International ed. in English) 2018, 57 (41), 13382–13392. https://doi.org/10.1002/anie.201802288. 2. Caschera, F.; Noireaux, V. Integration of Biological Parts toward the Synthesis of a Minimal Cell. Current opinion

in chemical biology 2014, 22, 85–91. https://doi.org/10.1016/j.cbpa.2014.09.028.

3. Exterkate, M.; Driessen, A. J. M. Synthetic Minimal Cell: Self-Reproduction of the Boundary Layer. ACS Omega

2019, 4 (3). https://doi.org/10.1021/acsomega.8b02955.

4. Dowhan, W.; Vitrac, H.; Bogdanov, M. Lipid-Assisted Membrane Protein Folding and Topogenesis. The protein journal 2019. https://doi.org/10.1007/s10930-019-09826-7.

5. Dowhan, W. Molecular Basis for Membrane Phospholipid Diversity: Why Are There so Many Lipids? Annual Review of Biochemistry 1997, 66, 199–232. https://doi.org/10.1146/annurev.biochem.66.1.199.

6. Hiraoka, S.; Matsuzaki, H.; Shibuya, I. Active Increase in Cardiolipin Synthesis in the Stationary Growth Phase and Its Physiological Significance in Escherichia Coli. FEBS letters 1993, 336 (2), 221–224.

7. Exterkate, M.; Caforio, A.; Stuart, M. C. A.; Driessen, A. J. M. Growing Membranes In Vitro by Continuous Phospholipid Biosynthesis from Free Fatty Acids. ACS synthetic biology 2018, 7 (1), 153–165. https://doi. org/10.1021/acssynbio.7b00265.

8. Arias-Cartin, R.; Grimaldi, S.; Pommier, J.; Lanciano, P.; Schaefer, C.; Arnoux, P.; Giordano, G.; Guigliarelli, B.; Magalon, A. Cardiolipin-Based Respiratory Complex Activation in Bacteria. Proceedings of the National Academy of Sciences of the United States of America 2011, 108 (19), 7781–7786. https://doi.org/10.1073/ pnas.1010427108 [doi].

9. Carranza, G.; Angius, F.; Ilioaia, O.; Solgadi, A.; Miroux, B.; Arechaga, I. Cardiolipin Plays an Essential Role in the Formation of Intracellular Membranes in Escherichia Coli. Biochimica et biophysica acta 2017, 1859 (6), 1124–1132. https://doi.org/S0005-2736(17)30080-9 [pii].

10. Lopalco, P.; Lobasso, S.; Babudri, F.; Corcelli, A. Osmotic Shock Stimulates de Novo Synthesis of Two Cardiolipins in an Extreme Halophilic Archaeon. Journal of lipid research 2004, 45 (1), 194–201. https://doi.org/10.1194/jlr. M300329-JLR200 [doi].

11. Romantsov, T.; Guan, Z.; Wood, J. M. Cardiolipin and the Osmotic Stress Responses of Bacteria. Biochimica et biophysica acta 2009, 1788 (10), 2092–2100. https://doi.org/10.1016/j.bbamem.2009.06.010 [doi].

12. Tan, B. K.; Bogdanov, M.; Zhao, J.; Dowhan, W.; Raetz, C. R. H.; Guan, Z. Discovery of a Cardiolipin Synthase Utilizing Phosphatidylethanolamine and Phosphatidylglycerol as Substrates. Proceedings of the National Academy of Sciences 2012, 109 (41), 16504–16509. https://doi.org/10.1073/pnas.1212797109.

13. Kawai, F.; Shoda, M.; Harashima, R.; Sadaie, Y.; Hara, H.; Matsumoto, K. Cardiolipin Domains in Bacillus Subtilis Marburg Membranes. Journal of Bacteriology 2004, 186 (5), 1475–1483.

14. Mileykovskaya, E.; Dowhan, W. Visualization of Phospholipid Domains in Escherichia Coli by Using the Cardiolipin-Specific Fluorescent Dye 10-N-Nonyl Acridine Orange. Journal of Bacteriology 2000, 182 (4), 1172– 1175.

15. Mileykovskaya, E.; Dowhan, W. Cardiolipin Membrane Domains in Prokaryotes and Eukaryotes. Biochimica et

3

(20)

biophysica acta 2009, 1788 (10), 2084–2091. https://doi.org/10.1016/j.bbamem.2009.04.003 [doi].

16. Nishijima, S.; Asami, Y.; Uetake, N.; Yamagoe, S.; Ohta, A.; Shibuya, I. Disruption of the Escherichia Coli Cls Gene Responsible for Cardiolipin Synthesis. Journal of Bacteriology 1988, 170 (2), 775–780.

17. Guo, D.; Tropp, B. E. A Second Escherichia Coli Protein with CL Synthase Activity. Biochimica et biophysica acta

2000, 1483 (2), 263–274. https://doi.org/S1388-1981(99)00193-6 [pii].

18. Li, C.; Tan, B. K.; Zhao, J.; Guan, Z. In Vivo and in Vitro Synthesis of Phosphatidylglycerol by an Escherichia Coli Cardiolipin Synthase. The Journal of biological chemistry 2016, 291 (48), 25144–25153. https://doi.org/ M116.762070 [pii].

19. Pluschke, G.; Hirota, Y.; Overath, P. Function of Phospholipids in Escherichia Coli. Characterization of a Mutant Deficient in Cardiolipin Synthesis. The Journal of biological chemistry 1978, 253 (14), 5048–5055.

20. Tropp, B. E. Cardiolipin Synthase from Escherichia Coli. Biochimica et biophysica acta 1997, 1348 (1–2), 192– 200. https://doi.org/S0005-2760(97)00100-8 [pii].

21. Ohta, A.; Obara, T.; Asami, Y.; Shibuya, I. Molecular Cloning of the Cls Gene Responsible for Cardiolipin Synthesis in Escherichia Coli and Phenotypic Consequences of Its Amplification. Journal of bacteriology 1985, 163 (2), 506–514.

22. Ivanisevic, R.; Milic, M.; Ajdic, D.; Rakonjac, J.; Savic, D. J. Nucleotide Sequence, Mutational Analysis, Transcriptional Start Site, and Product Analysis of Nov, the Gene Which Affects Escherichia Coli K-12 Resistance to the Gyrase Inhibitor Novobiocin. Journal of bacteriology 1995, 177 (7), 1766–1771. https://doi.org/10.1128/ jb.177.7.1766-1771.1995.

23. Quigley, B. R.; Tropp, B. E. E. Coli Cardiolipin Synthase: Function of N-Terminal Conserved Residues. Biochimica et biophysica acta 2009, 1788 (10), 2107–2113. https://doi.org/10.1016/j.bbamem.2009.03.016 [doi].

24. Shibuya, I.; Yamagoe, S.; Miyazaki, C.; Matsuzaki, H.; Ohta, A. Biosynthesis of Novel Acidic Phospholipid Analogs in Escherichia Coli. Journal of Bacteriology 1985, 161 (2), 473–477.

25. Caforio, A.; Driessen, A. J. M. Archaeal Phospholipids: Structural Properties and Biosynthesis. Biochimica et biophysica acta 2017, 1862 (11), 1325–1339. https://doi.org/S1388-1981(16)30343-2 [pii].

26. Jeucken, A.; Helms, J. B.; Brouwers, J. F. Cardiolipin Synthases of Escherichia Coli Have Phospholipid Class Specific Phospholipase D Activity Dependent on Endogenous and Foreign Phospholipids. Biochimica et biophysica acta 2018. https://doi.org/S1388-1981(18)30140-9 [pii].

27. Caforio, A.; Jain, S.; Fodran, P.; Siliakus, M.; Minnaard, A. J.; van der Oost, J.; Driessen, A. J. Formation of the Ether Lipids Archaetidylglycerol and Archaetidylethanolamine in Escherichia Coli. The Biochemical journal

2015, 470 (3), 343–355. https://doi.org/10.1042/BJ20150626 [doi].

28. Jain, S.; Caforio, A.; Fodran, P.; Lolkema, J. S.; Minnaard, A. J.; Driessen, A. J. M. Identification of CDP-Archaeol Synthase, a Missing Link of Ether Lipid Biosynthesis in Archaea. Chemistry & biology 2014, 21 (10), 1392–1401. https://doi.org/10.1016/j.chembiol.2014.07.022.

3

(21)

SUPPLEMENTARY MATERIALS

A

ClsA (Δ2-60) ClsA EV

B.1

Figure S1. A B.1

3

(22)

B.2

Figure S1. (a) Western blot analysis of a purified empty vector control, ClsA and ClsA(Δ2-60) using a

his-tag antibody. (b) Proteomic analysis of (b.1) ClsA and (b.2) ClsA(Δ2-60) after in gel trypsin digestion and subsequent LC-MS analysis.

B.2

3

(23)

Figure S2. In vitro ClsA activity in the presence of DOPG and POPA. Lipid species were analyzed by

LC-MS, normalized for the internal standard and plotted.

Control ClsA ClsA + 2,2-Me-1,3-prOH ClsA + 2-Phe-1,3-prOH 0.0 0.2 0.4 0.6 0.8 1.0 Lipid species (norma liz ed ion count) PG CL PA p-2,2-Me-1,3-prOH p-2-Phe-1,3-prOH

Control ClsA ClsA + amino-propanol ClsA + ethanol-amine 0.0 0.1 0.2 0.3 0.4 0.5 Lipid species (norma liz ed ion count) CL PG PA p-NH2-prOH PE Control ClsA 0.0 0.2 0.4 0.6 0.8 1.0 Lipid species (normaliz ed ion count) PG CL PA p-1-NH2-prOH PE ClsA + amino- ethanol-ClsA +

A

B

C

Control ClsA ClsA + amino-propanol ClsA + ethanol-amine 0.0 0.1 0.2 0.3 0.4 0.5 Lipid species (norma liz ed ion count) CL PG PA p-NH2-prOH PE Control ClsA 0.0 0.2 0.4 0.6 0.8 1.0 Lipid species (normaliz ed ion count) PG CL PA p-1-NH2-prOH PE ClsA + amino-propanol ClsA + ethanol-amine

A

B

C

Control ClsA ClsA + amino-propanol ClsA + ethanol-amine 0.0 0.1 0.2 0.3 0.4 0.5 Lipid species (norma liz ed ion count) CL PG PA p-NH2-prOH PE Control ClsA 0.0 0.2 0.4 0.6 0.8 1.0 Lipid species (normaliz ed ion count) PG CL PA p-1-NH2-prOH PE ClsA + amino-propanol ClsA + ethanol-amine

A

B

C

B A C

Figure S3. In vitro ClsA activity in the presence of various alcohols and the substrate PG or CL. (a)

glycerol (1,2,3-propanetriol), 2,2-dimethyl-1,3-propanediol and 2-phenyl-1,3-propanediol. (b and c) 1-amino-3-propanol, ethanolamine. Lipid species were analyzed by LC-MS, if noted normalized for the internal standard and plotted.

(24)

A

B

Figure S4. (a) Schematic representation of the BLAST homology search on ClsA, revealing the two

main archaeal clusters containing halophilic archaea and methanogenic archaea respectively. (b) Coomassie stained SDS-PAGE gel of purified Mh_Cls by Ni-NTA chromatography.

O O O OH P OH OH O O O O O OH P OH OH O O also synthesized 14' - sn-1 - sn-3 diastereomer

14 - sn-1, sn-1 diastereomer

Figure S5. Schematic representation of the synthesized racemic archaetidylglycerol AG.

3

(25)