Engineering Acyl Carrier Protein to Enhance Production of Shortened Fatty Acids

Xueliang Liu1,2# Wade M. Hicks1,3#*, Pamela A. Silver1,3, and Jeffrey C. Way1*

1 Wyss Institute for Biologically Inspired Engineering; 2 School of Engineering and Applied

Sciences, Harvard University; 3 Department of Systems Biology, Harvard Medical School

Keywords: acyl carrier protein; protein engineering; thioesterase; free fatty acid; biofuels

Abstract

We engineered the specificity of the acyl-carrier protein (ACP), an essential enzyme of fatty acid metabolism, to alter the E. coli lipid pool and enhance production of medium-chain fatty acids as biofuels precursors. A homology model of the S. elongatus ACP was constructed showing a hydrophobic pocket harboring the growing acyl chain. Amino acids within the pocket were mutated to increase steric hindrance to the acyl-chain, and gas chromatography-mass spectrometry analysis performed on cellular extracts identified mutants that, when over- expressed in E. coli, increased the proportion of shorter-chain lipids; I75W and I75Y showed the strongest effects. We further demonstrated increased production of lauric acid in E. coli also expressing the C12-specific acyl-ACP thioesterase from Cuphea palustris.

Introduction

With the continuous rise in global energy needs and adverse climate changes, development of cleaner and renewable alternatives to fossil fuels has become paramount. Microbial synthesis of biofuels is an attractive, renewable alternative to fossil fuels.1â??3
Organisms naturally synthesize large quantities of fuel-like hydrocarbons in the form of lipids,
which are used in cell membranes and other molecules. In microbes, the end products of fatty acid metabolism are long acyl-chains consisting mostly of 16-18 carbons. When extracted for fuels, these long chain carbon molecules remain solid at room temperature and lack favorable physical properties such as higher volatility and lower viscosity. Such properties are characteristic of medium-length (8-12) carbon chains used ubiquitously in fuels for vehicles and jets.
Previous work on the biological synthesis of medium length fuel precursors has employed thioesterase enzymes with medium-length chain specificity to release free fatty acids (FFA) from intermediates in fatty acid synthesis.4â??7 Here we employ a complementary strategy
to bias FFA synthesis toward shorter chains by engineering acyl-carrier protein (ACP), an essential enzyme and key component of fatty acid metabolism. In fatty acid synthesis in bacteria and plants, ACP is attached to the acyl chain and presents it to the other enzymes during successive cycles of elongation and reduction (Figure 1).8â??11 ACP is a small (~9kB), acidic (pI =
4.1) protein abundant in the cytoplasm, constituting about 0.25% of all soluble proteins in E.
coli.8 The structure of ACP is highly conserved even among variants with low sequence similarity. There are, in general, four alpha helices with the major helices I, II and IV running parallel to each other enclosing a hydrophobic pocket that harbors the acyl-chain and minor helix III running perpendicular to these (Figure 2). The acyl chain is connected to a 4- phosphopantetheine modification at a conserved serine and enters the hydrophobic cavity between helices II and III. Roujeinikova et al. solved the structures of E. coli ACP attached to C6, C7, and C10 fatty acids.12 In each case, the distal end of the fatty acid terminates in a deep pocket within the protein near Ile72 (corresponding to Ile75 of the S. elongatus ACP), with the phosphopantetheine group also entering the pocket to varying degrees. Acyl chains up to 8 carbons are fully bound within the pocket, with the thioester bond sequestered in the core of the protein.8,12â??14 We therefore hypothesized that the size of ACPâ??s hydrophobic pocket serves as a molecular ruler that influences the composition of lipid lengths in a cell. As the acyl chain grows to a length of around 16, the thioester bond becomes fully solvent exposed, which may facilitate cleavage by phosphate:acyl-ACP acyltransferases for downstream phospholipid processing.
To test our molecular ruler hypothesis, we have mutated residues at the base of the hydrophobic pocket to contain bulkier hydrophobic side chains. We found that over-expressing certain mutant ACPs increased production of medium chain fatty acids. Our findings could be potentially advantageous for microbial production of transportation biofuels based on artificial metabolically engineered pathways.

Results and Discussion

To test our hypothesis of ACP acting as a molecular ruler for growing acyl-chains, we constructed mutants of ACP designed to decrease the acyl-chain pocket size (Figure 2). Variants of the cyanobacterial (S. elongatus) ACP were expressed in an E. coli host. We chose S. elongatus ACP due to its potential compatibility with recently discovered enzymes of the cyanobacterial alkane biosynthesis pathway15, which could enable microbial synthesis of fatty alcohol or alkanes. To determine which hydrophobic residues of S. elongatus ACP lined the inner, acyl-chain pocket, we constructed a structural homology model using the published crystal structure of E. coli ACP bound to a C10 fatty acyl chain (2FAE) as a template (Figure 2). We constructed a number of single amino acid mutants by exchanging small hydrophobic side chain residues, such as isoleucine or leucine, with bulkier hydrophobic side-chains such as phenylalanine, methionine, tyrosine or tryptophan. ACPs initially fold into an inactive apo state. Conversion to the active holo state is achieved through post-translational modification whereby
4â??-phosphopantetheine is transferred from co-enzyme A (CoA) to a specific serine residue on the apo-ACP (Ser39 on S. elongatus ACP).8,16 ACP over-expression may reduce the CoA pool leading to toxic accumulation of apo-ACP, which inhibits sn-glycerol-3-phosphate acyltransferase lipid metabolism and suppresses cellular growth.16,17 To verify functional expression of recombinant ACPs, we measured culture growth kinetics over 15 hours. Compared to controls, cells overexpressing WT E. coli ACP (Ec-ACP), WT S. elongatus ACP (Se-ACP), or mutant Se-ACPs all showed suppressed growth at low levels of induction and worsened at higher induction levels (Figure 3). The decrease in culture density after 15 hours was as large or larger for Se-ACP mutants than for WT Se-ACP (Figure S1). Growth suppression by the point mutants suggests that these recombinant cyanobacterial ACPs were expressed and properly folded.
To analyze the effect of mutant Se-ACPs on lipid pools, we used gas chromatography mass spec (GC-MS) to characterize fatty acid methyl esters (FAMEs) derived from lipid pools in
Se-ACP overexpressing cells. We compared ratios of FAME peak areas for each sample to remove the influence of growth kinetics and cell densities. We detected peaks for FAMEs derived from the naturally most abundant palmitic acid (C16) and the shorter, less abundant myristic acid (C14) and quantified these peaks in all sample spectra. The ratios of C14 and C16 peak areas were calculated and compared across controls and Se-ACP point mutants. For all uninduced samples, the C14:C16 ratio is around 0.1 (Figure 4A). After induction, only the I75W and I75Y Se-ACP mutants demonstrated a statistically significant increase in the C14:C16 ratio relative to WT Se-ACP, 3- and 2.7-fold respectively (p<0.05, two-tailed student-t test) (Figure
4B), indicating that their lipid pools had shifted toward shorter acyl chains. Mutants that replaced Leu49 or Ile57 did not increase the proportions of shorter fatty acids compared to over- expressing WT ACPs. The side chain of isoleucine 75 is positioned in the hydrophobic pocket close to the terminus of the acyl-chain, more so than residues 49 and 57 (Figure 2A).12 We suggest that mutations at position 75 more greatly enhance steric hindrance inside the ACP pocket (Figure 2B and 2C). The I75W and I75Y mutations seem sufficient to bias the fatty acid pool toward a higher proportion of shorter chains, however, mutating Ile75 to phenylalanine or methionine did not cause nearly as significant shifts in lipid pool chain-length composition (Figure 4). This could be because tyrosine and tryptophan have bulkier hydrophobic side chains compared to methionine and phenylalanine and therefore create a smaller acyl-chain pocket within ACP. Indeed, our homology models indicate that the Tyr75 and Trp75 sidechains protrude roughly two carbon-carbon bond distances further into the hydrophobic acyl-chain pocket than an isoleucine at the same position (Figure 2B and 2C; only I75W shown). Therefore, I75W and I75Y Se-ACP mutants may directly hinder elongation from C14 to C16 in fatty acid synthesis and skew the fatty acid pool towards shorter chain lengths. C16 lipids are still dominant, possibly due to the presence of WT Ec-ACP, and/or selectivity at the level of phospholipid formation, or an inherent preference of ACP for C16 and C18 chain lengths being altered to C14 and C16. A certain level of C16 lipids may be necessary for cell membrane
integrity and survival. This may also explain the stronger growth defect found in I75W and I75Y mutant Se-ACP strains compared to the control strain overexpressing WT Se-ACP (Figure 3, Figure S1).
To explore the potential to further skew cellular lipids towards short chain lengths, particular those shorter than 14 carbons long, we introduced secondary point-mutations in addition to the Se-ACP I75W or I75Y mutations (Table 1). Here amino acids with small hydrophobic side-chains such as isoleucine, valine or alanine were exchanged for a bulkier methionine, a polar glutamine, or a hydrophilic arginine and assayed lipid content for changes to the C14:C16 ratio. Our results show that double mutant Se-ACPs do not significantly increase the C14:C16 ratio beyond either single I75W or I75Y mutation alone (Figure S2). Chains shorter than C14 were not detected at reliable quantities by GC-MS most likely due to their naturally low abundance as transient intermediates in FFA synthesis rather than accumulated end products in membranes.
As an additional control, Se-ACP serine 39 residue, which is post-translationally modified with 4-phosphopantethene, was mutated to alanine (S39A), thereby generating an inactive, obligate apo-ACP. Overexpressing this inactive ACP resulted in similarly low C14:C16 ratio compared to WT (Figure 4). Furthermore, cellular growth was strongly suppressed by over- expressing this mutant, confirming that an overabundance of apo-ACP induces toxicity.
A 24 hour time-course measuring lipid pool composition shows that the highest C14:C16 ratio occurs around 5 hours post-induction (Figure 5). Longer induction times result in a decreased C14:C16 ratio for all strains, particularly for Se-ACP I75W and I75Y mutants, which become indistinguishable from WT controls by 24 hours. This highlights the importance of growth phase on lipid composition. During exponential growth phase when cells are actively dividing and building new membranes, fatty acid metabolism is highly active, and an abundance
of mutated ACPs with reduce pocket sizes likely biases the fatty acid pool toward shorter acyl- chains. It may be that as cell growth slows membrane synthesis proceeds with greater fidelity. Alternatively, short-chain fatty acids may be degraded or â??repairedâ?? and actively replaced with fatty acids of the correct length, which would be more apparent in stationary phase when new C14 fatty acids are not being added to membrane lipids.
We next tested the effect of mutant ACPs on production of shorter chain FFAs, fatty alcohols and alkanes. To over-produce FFAs, a thioesterase that specifically produces 12- carbon chains (CpFatB1 from Cuphea palustris)6 was co-expressed with Se-ACP. We hypothesized that by expressing our mutant ACPs, increased cellular concentration of shorter chain acyl-ACPs would serve as substrates to the medium chain-specific thioesterase and further increase the yield of medium chain FFAs. To test this hypothesis, we expressed the C12-specific thioesterase along with WT or mutant Se-ACP and measured FFA production by GC-MS analysis of fatty acid ethyl esters (FAEE) derived from produced FFAs (Figure 6).
In conjunction with expressing the C12 thioesterase, strains over-expressing I75W or I75Y mutant ACPs significantly increased medium chain FFA yields compared to those expressing either WT Ec- or Se-ACP (Figure 6). Double mutants based on I75W or I75Y (e.g. I75W+I57M) also produced similar amounts of medium chain FFAs as the single mutants alone. Similar to our hypothesis on ACP functionality, C12-specific and other medium-chain thioesterases likely act as molecular rulers specifically recognizing fatty acid chains extruded from ACP. We propose that overexpressing mutant Se-ACPs enhances thioesterase-dependent C12 FFA production simply by increasing the C12-acyl-ACP substrate pool. Empty vector and obligate apo-ACP (S39A) controls both produced less FFA than the I75W or I75Y mutants but more than the WT Ec- or Se-ACP controls. Over-expression of active holo-ACPs could burden cells by depleting CoA pools through excess 4-phosphopantetheine modification of ACP.16,17 If
so, we may be underestimating the enhanced medium-chain FFA synthesis observed for the
I75W and I75Y mutants.
We have shown that ACP, an essential enzyme in fatty acid metabolism, can be modified by site-directed mutagenesis to skew cellular lipid pools towards smaller acyl-chain lengths. Specifically, expressing mutant ACPs enhanced the level of C14 fatty acids in membrane lipids, and by co-expressing mutant ACPs with chain-length specific thioesterases we have demonstrated the ability to tune the length of fatty acid production. Taken together, these results support our hypothesis that bacterial ACPs function as a molecular ruler during fatty acid synthesis. Other enzymes involved in fatty acid synthesis also likely function as molecular rulers and engineering modified acyl-chain specificity has been similarly achieved. For example, FabB and FabF catalyze elongation of fatty acid chains (Figure 1), and have a clearly defined pocket that should accommodate carbon chains up to about 18.18 Val et al.
engineered the FabF pocket to accommodate a maximum of 6 carbons.19 Similarly, the
cyanobacterial aldehyde decarbonylase solved structure20,21 contains electron density corresponding to a C18 fatty acid or aldehyde; Khara et al. modified this enzyme to have specificity for medium-chain substrates. The C8, C12 and C14-specific plant-derived Acyl-ACP thioesterases apparently also function as molecular rulers, although the underlying structural mechanisms have not been identified. Since FFAs are hydrophilic, they are not ideal fuel molecules. Instead, FFAs can act as precursors to further enzymatic modification for in transformation into highly desired fuel molecules such as fatty alcohols and alkanes. Engineering such enzymes (e.g. aldehyde decarbonylases, acyl-ACP reductases, and carboxylic acid reductases) towards shorter carbon chain substrate recognition will likely be key to tailoring biofuel formulations. To achieve the ultimate goal of efficient biofuel synthesis, it may be necessary to engineer the length specificity of several enzymes â?? most such enzymes have evolved to handle chains of 16-18 carbons, but shorter chains are desired in fuels. This
technology could help to optimize biofuel yield and molecular makeup, which would benefit the goal of developing energy sources alternative to fossil fuels.

Methods

Homology Modelling

The structural model of Se-ACP harboring a decanoyl-chain was obtained by homology to the published x-ray crystal structure of the E. coli decanoyl-ACP (2FAE) using SWISS-MODEL.12

Strain Construction

Double stranded DNA encoding E. coli and S. elongatus ACP genes were synthesized as gBlocks (Integrated DNA Technologies) and cloned into the pCDF-Duet vector by Gibson Assembly.22 Single and double amino acid mutations of the Se-ACP gene were incorporated during DNA synthesis. An empty pCDF-Duet-1 vector (Millipore) without the ACP gene was included as control. Plasmids were sequence-verified and transformed into E.coli BL21(DE3). For free fatty acid (FFA) production, the C12 thioesterase gene (CpFatB1 from Cuphea palustris) was cloned into pET-Duet-1 vector (Millipore) and transformed into strains harboring the plasmids carrying the ACP variants. For fatty alcohol and alkane production, in one pathway the S. elongatus acyl-ACP reductase and aldehyde decarbonyase were cloned into pCOLA-Duet-1 and the pACYC-Duet-1 vectors, respectively. The plasmids were transformed into the strains expressing ACP (in the pCDF-Duet-1 vector) by electroporation. In an alternative pathway, a carboxylic acid reductase (Mycobacterium marinum), its maturation factor phosphopantetheinyl transferase (Bacillus subtilis), and the petF ferredoxin gene were cloned into the pET-Duet vector with the C12 thioesterase. An aldehyde decarbonylase (Prochlorococcus marinus) was cloned into pRSF-Duet-1 vector. The pET-Duet and pRSF-Duet vectors were transformed into the ACP-expressing strains by electroporation. The sequences of these plasmids are given in Supplementary Materials.

Growth Kinetics Assay

ACP expressing strains in triplicates were inoculated from single colonies representing independent transformants into LB medium, grown overnight to saturation and back-diluted into M9 minimal media containing 0.4% glucose. The cultures were grown to mid-exponential phase (OD~0.4), dispersed into 96-well plates, induced with various concentrations of IPTG and left to grow shaking at 37Ë?C in a plate reader (BioTek NEO). The optical densities (OD) of the cultures were recorded every 5 minutes over 15 hours by the plate reader. The growth curves, as well as the final OD after 15 hours were compared among the strains to quantify growth suppression by ACP over-expression.

Analysis of Cellular Lipid Composition

ACP expressing strains in triplicates were inoculated in LB and grown overnight and back- diluted into M9 minimal media containing 3% glucose. The cultures were grown to an optical density of 0.4, induced with 1mM IPTG, and grown for 6 more hours at 37Ë?C. For the time course experiment, the cultures were left to grow for up to 24 hours. After growth, the cells were pelleted and resuspended in 1:1 methanol:chloroform with 2% glacial acetic acid for lysis, hydrolysis of membrane lipids, and solubilization of fatty acids into the organic phase. An octanoate (C8 fatty acid) was added into the mixture as an internal standard. After vigorous mixing by vortexing, the organic phase was transferred by glass pipettes into glass vials, and the chloroform solvent was evaporated by nitrogen. The vials were then treated with methanol containing 1.25M HCl at 50 50Ë?C for 15 hours to catalyze methylation of the fatty acids. The reaction was quenched by adding 5ml of 100mg/ml sodium bicarbonate. 0.5ml hexane was added and the mixture was vortexed vigorously before the hexane phase containing the fatty acid methyl esters (FAME) was extracted and subsequently analyzed on a GC-MS (Agilent
6890/5975).23 First a standard set of FAMEs with varying chain lengths was run on the GC-MS
in scan mode to determine the identity of each fatty acid peak based on the elution time for each fatty acid and comparison of its fragment profile to those in the NIST database (via Agilent ChemStation software). Fatty acid peaks from the extracted cell samples were also identified using scan mode. To quantify peak areas, the background was minimized by using Selective Ion Mode (SIM) whereby the elution times were used to determine fatty acid identity and only the most dominant mass peaks pertaining to each fatty acid methyl ester were counted. To compare the proportions of different chain lengths in each sample, the ratio of the peak areas for 14-carbon chain to 16-carbon chain was taken.

Analysis of Free Fatty Acid (FFA)

ACP and C12 thioesterase expressing strains in triplicates were grown in M9 minimal media containing 3% glucose and induced with IPTG as described above. After 6 or 24 hours of growth, 5 microliters of each culture (cells and media, as medium chain FFA may be secreted) was transferred to wells of a new 96-well plate for high-throughput spectrometric determination of FFA concentration using the Roche Free Fatty Acid Kit (Product Number 11383175001). To specifically quantify lauric acid, cultures of ACP plus thioesterase-expressing cells were lysed and extracted with chloroform. The FFA was ethylated and run on the GC-MS to determine the spectrum of chain lengths.

Author Information

Corresponding Authors

*EMAIL: wade.hicks@wyss.harvard.edu
*EMAIL: jeff.way@wyss.harvard.edu

Author Contributions

# XL and WMH contributed equally to this work. References

(1) Lennen, R. M., and Pfleger, B. F. (2013) Microbial production of fatty acid-derived fuels and chemicals. Curr. Opin. Biotechnol. 1â??10.
(2) Choi, Y. J., and Lee, S. Y. (2013) Microbial production of short-chain alkanes. Nature. (3) Dellomonaco, C., Clomburg, J. M., Miller, E. N., and Gonzalez, R. (2011) Engineered
reversal of the β-oxidation cycle for the synthesis of fuels and chemicals. Nature 476, 355â??9.
(4) Zheng, Y.-N., Li, L.-L., Liu, Q., Yang, J.-M., Wang, X.-W., Liu, W., Xu, X., Liu, H., Zhao, G., and Xian, M. (2012) Optimization of fatty alcohol biosynthesis pathway for selectively enhanced production of C12/14 and C16/18 fatty alcohols in engineered Escherichia coli. Microb. Cell Fact.

11, 65.

(5) Lennen, R. M., and Pfleger, B. F. (2012) Engineering Escherichia coli to synthesize free fatty acids. Trends Biotechnol. 30, 659â??67.
(6) Torella, J. P., Ford, T. J., Kim, S. N., Chen, A. M., Way, J. C., and Silver, P. a. (2013)
Tailored fatty acid synthesis via dynamic control of fatty acid elongation. Proc. Natl. Acad. Sci. U. S. A. 110, 11290â??5.
(7) Voelker, T. A., and Davies, H. M. (1994) Alteration of the specificity and regulation of Alteration of the Specificity and Regulation of Fatty Acid Synthesis of Escherichia coli by Expression of a Plant Medium- Chain Acyl-Acyl Carrier Protein Thioesterase.
(8) Chan, D. I., and Vogel, H. J. (2010) Current understanding of fatty acid biosynthesis and the acyl carrier protein. Biochem. J. 430, 1â??19.
(9) Crosby, J., and Crump, M. P. (2012) The structural role of the carrier protein â?? active controller or passive carrier â?  1111â??1137.
(10) Nguyen, C., Haushalter, R. W., Lee, D. J., Markwick, P. R. L., Bruegger, J., Caldara-Festin, G., Finzel, K., Jackson, D. R., Ishikawa, F., Oâ??Dowd, B., McCammon, J. A., Opella, S. J., Tsai, S.-C., and Burkart, M. D. (2014) Trapping the dynamic acyl carrier protein in fatty acid biosynthesis. Nature 505, 427â??31.
(11) Masoudi, A., Raetz, C. R. H., Zhou, P., and Pemble IV, C. W. (2013) Chasing acyl carrier protein through a catalytic cycle of lipid A production. Nature 1â??7.
(12) Roujeinikova, A., Simon, W. J., Gilroy, J., Rice, D. W., Rafferty, J. B., and Slabas, A. R. (2007) Structural studies of fatty acyl-(acyl carrier protein) thioesters reveal a hydrophobic binding cavity that can expand to fit longer substrates. J. Mol. Biol. 365, 135â??45.
(13) Chan, D. I., Stockner, T., Tieleman, D. P., and Vogel, H. J. (2008) Molecular dynamics simulations of the Apo-, Holo-, and acyl-forms of Escherichia coli acyl carrier protein. J. Biol. Chem. 283, 33620â??9.
(14) Roujeinikova, A., Baldock, C., Simon, W. J., Gilroy, J., Baker, P. J., Stuitje, A. R., Rice, D. W., Slabas, A. R., Rafferty, J. B., and Sheffield, S. (2002) X-Ray Crystallographic Studies on Butyryl-ACP Reveal Flexibility of the Structure around a Putative Acyl Chain Binding Site 10,
825â??835.
(15) Schirmer, A., Rude, M. a, Li, X., Popova, E., and del Cardayre, S. B. (2010) Microbial biosynthesis of alkanes. Science 329, 559â??62.
(16) Carey, M. R. (1995) The Unmodified (Apo) Form of Escherichia coli Acyl Carrier Protein Is a Potent Inhibitor of Cell Growth. J. Biol. Chem. 270, 22229â??22235.
(17) Magnuson, K., Jackowski, S., Rock, C. O., and Cronan, J. E. (1993) Regulation of fatty acid biosynthesis in Escherichia coli. Microbiol. Rev. 57, 522â??42.
(18) White, S. W., Zheng, J., Zhang, Y.-M., and Rock. (2005) The structural biology of type II
fatty acid biosynthesis. Annu. Rev. Biochem. 74, 791â??831.
(19) Val, D., Banu, G., Seshadri, K., Lindqvist, Y., and Dehesh, K. (2000) Re-engineering ketoacyl synthase specificity. Structure 8, 565â??566.
(20) Jia, C., Li, M., Li, J., Zhang, J., Zhang, H., Cao, P., Pan, X., Lu, X., and Chang, W. (2015) Structural insights into the catalytic mechanism of aldehyde-deformylating oxygenases. Protein Cell 6, 55â??67.
(21) Khara, B., Menon, N., Levy, C., Mansell, D., Das, D., Marsh, E. N. G., Leys, D., and Scrutton, N. S. (2013) Production of propane and other short-chain alkanes by structure-based engineering of ligand specificity in aldehyde-deformylating oxygenase. Chembiochem 14, 1204â??
8.
(22) Gibson, D. G., Young, L., Chuang, R.-Y., Venter, J. C., Hutchison, C. a, and Smith, H. O. (2009) Enzymatic assembly of DNA molecules up to several hundred kilobases. Nat. Methods 6,
343â??5.
(23) Lennen, R. M., Braden, D. J., West, R. a, Dumesic, J. a, and Pfleger, B. F. (2010) A process for microbial hydrocarbon synthesis: Overproduction of fatty acids in Escherichia coli and catalytic conversion to alkanes. Biotechnol. Bioeng. 106, 193â??202.

Figures

Figure 1
Membrane phospholipid

S O ACP

O O

Free
C12

PlsB

S

FabB FabF

S

Fatty
Acid

ACP

thioesterase

FabI

1+2n

O

ACP

O O

FabG

1+2n

ACP

S

1+2n

O

FabA FabZ

ACP

S

O OH

1+2n

Figure 1 Overview of Fatty acid Synthesis. Fatty acid synthesis proceeds through iterative cycles of elongation. In each cycle, the acyl-chain is extended by 2-carbons using a malonyl- ACP as a carbon donor (by FabB or FabF) and subsequently reduced into a saturated chain (by FabG, FabA, FabZ and FabI). From the first 2-carbon malonyl-ACP to the final length fatty acid processed through this cycle, the hydrophobic acyl-chain is attached to and shielded by the ACP instead of existing in a free form.

Figure 2

Figure 2 Se-ACP Structural Homology Models with WT and Mutant Residues. A. Homology model of Se-ACP bound to a C10 acyl-chain is shown. Highlighted in blue (residue 49), green (residue 57) and red (residue 75) are small hydrophobic amino acids lining the WT ACP pocket, Leu, Ile, and Ile, respectively. Each residue was separately mutated to a bulkier hydrophobic amino acid: methionine, tyrosine or tryptophan in order to induce steric hindrance and favor shorter-chain fatty acid synthesis. B. For simplicity, a homology model with all three residues of interest mutated to tryptophan shows how each side chain might be positioned when mutated separately. Trp75 (red) extends closest to the acyl-chain terminus. C. Looking up through the axis of the acyl chain from the bottom perspective of the ACP pocket, Trp75 (red) is more directly in line with the acyl chain, as compared to the other mutant residues. This positioning likely introduces the most direct steric hindrance to the acyl chain, resulting in increased production of shorter chain lipids.

Figure 3

Figure 3 Growth Suppression by ACP expression. Growth of E. coli is suppressed by induction of Se-ACP expression increasing from 0mM (black), 0.5mM (blue) to saturation at
1mM (red) IPTG. The growth defect is likely due to inhibition of phospholipid metabolism by apo-ACP. Se-ACP I75W expression (B) shows similar growth suppression compared to WT (A), indicating proper folding and functionality of ACP. All mutant ACPs show similar growth curves (data not shown). Representative growth curves are shown.
Figure 4

Figure 4 GC-MS analysis of Cellular Lipids in Single ACP Mutants. A. Ratios of C14 to C16

GC-MS peak areas for uninduced (black) and induced (red) strains. B. Fold changes of induced vs. uninduced C14:C16 ratios. The I75W and I75Y mutants have significantly increased C14:C16 ratios as compared to expressing WT Se-ACP. Data represents triplicate biological measurements. Error bars are standard error of the mean (S.E.M). * p<0.05, two-tailed student-t test.
Figure 5

Figure 5 Time Course of C14:C16 Ratios Se-ACP I75W and I75Y both demonstrate the highest C14:C16 cellular lipid ratio at 5 hours after induction during the growth phase. As the cell cultures saturate past 14 hours, the ratios decrease to the baseline of around 0.05-0.1. Data represent triplicate biological measurements. Error bars are S.EM.

Figure 6

Figure 6 Free Fatty Acid Production by C12 thioesterase. A. Representative GC-MS trace of FAEEs derived from cell cultures shows thioesterase specificity toward 12-carbon acyl-chains. B. FFA concentrations measured from cell cultures at 6 hours (blue) and 24 hours (black) post- induction of both the C12 thioesterase and the indicated ACP. The Se-ACP I75W and I75Y mutants and their derivatives yield more FFA than controls. WT Ec-ACP and Se-ACP controls yield the least FFA even compared to the empty vector (MT) and apo-ACP (S39A) controls. Data represent triplicate biological measurements. Error bars are S.E.M.

Supporting Information

Figure S1

Figure S1 Growth Suppression by ACP expression. Shown are changes in culture OD of E. coli strains induced to overexpress various ACPs versus their uninduced condition (0mM IPTG). Culture densities were measured after 15 hours of growth in M9 minimal media with 0.4% glucose. When overexpressed, most mutants show equal or stronger growth suppression vs. WT Se-ACP. Data represent triplicate biological measurements. Error bars are S.E.M.
Figure S2

Figure S2 GC-MS analysis of cellular lipids from double mutants. Combining the single Se- ACP I75W or I75Y point mutants with a second set of residues mutated to methionine does not significantly change the C14:C16 ratio from that observed for the single point mutants alone. Mutating this second set of residues to arginine (A71R) or glutamine (A71Q) reduced the C14:C16 ratio to WT Se-ACP levels. Data represent triplicate biological measurements. Error bars are S.E.M.