The initial step of the glycerolipid pathway: identification of glycerol-3-phosphate /

dihydroxyacetone phosphate dual substrate acyltransferases in Saccharomyces cerevisiae

Zhifu Zheng and Jitao Zou∗

Plant Biotechnology Institute, National Research Council of Canada, 110 Gymnasium Place,

Saskatoon, Sask., Canada S7N OW9

Running title: yeast G-3-P / DHAP acyltransferases

∗To whom correspondence should be addressed: Plant Biotechnology Institute, National

Research Council of Canada, 110 Gymnasium Place, Saskatoon, Sask., Canada S7N OW9

Phone: (306) 975 5583

Fax: (306) 975 4839

E-mail jitao.zou@nrc.ca

Keywords: Saccharomyces cerevisiae, glycerolipids, fatty acyltransferase, glycerol-3-phosphate

acyltransferase, dihydroxyacetone phosphate acyltransferase

1

Copyright 2001 by The American Society for Biochemistry and Molecular Biology, Inc.

JBC Papers in Press. Published on September 5, 2001 as Manuscript M104749200 by guest on N

ovember 30, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Summary

The initial step of phospholipid biosynthesis in yeast is carried out through the acylation of

glycerol-3-phosphate (G-3-P) and dihydroxyacetone phosphate (DHAP) by sterospecific sn-1

acyltransferases. Here we report the identification of two key fatty acyltransferases of the

glycerolipid biosynthesis pathway in Saccharomyces cerevisiae. Disruption of the open reading

frameYBL011w, corresponding to a gene previously identified as a choline transporter

suppressor (SCT1), resulted in a substantial decrease of total cellular G-3-P acyltransferase

activity. A yeast strain disrupted at the open reading frameYKR067w, which encodes a protein

closely related to Sct1p, also exhibited a dramatic reduction in G-3-P acyltransferase activity.

Molecular characterizations of the genes revealed that a missense mutation in YKR067w

accounted for a defect in the activities of the G-3-P acyltransferase in the yeast mutant strain

TTA1. Heterologous expression of YKR067w in E. coli further confirmed its enzyme activity.

These results indicate that YKR067w and YBL011w, designated herein as GAT1 and

GAT2(SCT1), respectively, are yeast G-3-P acyltransferase genes. Furthermore, biochemical results

are presented to show that both Gat1p and Gat2p(Sct1p) are G-3-P/DHAP dual substrate-

specific sn-1 acyltransferases. The fatty acyl specificity of Gat1p is similar to that of the

mammalian microsomal G-3-P acyltransferase, as it can effectively utilize a broad range of fatty

acids as acyl donors. In contrast, Gat2p(Sct1p) displayed preference towards 16-carbon fatty

acids. The most notable of the altered phospholipid compositions of the gat1∆ and gat2(sct1)∆

strains are a decreased phosphatidic acid (PA) pool and an increased phosphatidyl serine (PS) /

phosphatidyl inositol (PI) ratio. This did not appear to affect the mutants as no growth defect was

2

by guest on Novem

ber 30, 2020http://w

ww

.jbc.org/D

ownloaded from

found. However, null mutations of both GAT1 and GAT2(SCT1) are synthetically lethal to

yeast.

Studies on the new roles of phospholipids as structural elements in membranes, as well as

cell signaling components, continue to be at the forefront of our efforts in understanding a wide

range of biological processes (1-4). Meanwhile, there are still significant knowledge gaps with

regard to various aspects of regulation of the phospholipid biosynthetic pathway (1, 2). It is well

established that the initial step of phospholipid biosynthesis involves the acylation of G-3-P at

the sn-1 position by a G-3-P acyltransferase to form lysophosphatidic acid (LPA). LPA

acyltransferase then catalyzes the acylation of LPA at the sn-2 position to generate phosphatidic

acid (PA), which serves as a general precursor for all glycerophospholipids, including

triacylglycerol in eukaryotes (5, 6). In Escherichia coli, an integral membrane protein (plsB) is

responsible for the G-3-P acyltrasnferase activity, and its corresponding gene has been

identified (7). In eukaryotic cells, multiple isoforms of G-3-P acyltransferase are present and

localized in different intracellular compartments (8, 9). The genes corresponding to the

mammalian mitochondrial and plant plastidial G-3-P acyltrasnferase have been isolated and

characterized in detail (8, 9). In contrast, the eukaryotic microsomal counterpart has so far

remained elusive, mainly due to the difficulties encountered in the purification of these

membrane proteins and reconstitution of functional enzymes.

Baker’s yeast, Saccharomyces cerevisiae, is a convenient model organism for eukaryotic

lipid studies since its glycerolipid biosynthetic pathway is very similar to a wide range of species

including higher plants and mammals. More than a decade ago, Tillman and Bell reported

mutants deficient in the activities of G-3-P acyltransferase in S. cerevisiae (10). Subsequent

3

by guest on Novem

ber 30, 2020http://w

ww

.jbc.org/D

ownloaded from

biochemical characterizations of one such mutant, generally known as TTA1, have yielded many

new insights into lipid metabolism. It is now widely accepted that the initial step of glycerolipid

biosynthesis in yeast is mediated by a G-3-P/DHAP dual substrate acyltransferase (10, 12), and

that multiple isoforms of G-3-P acyltransferase are present in yeast (11, 12). However, the

protein and the gene corresponding to the mutation have not been identified due to the lack of an

apparent selectable growth phenotype in TTA1.

Based on our general interest in eukaryotic fatty acyltransferases, we pursued the

identification of G-3-P acyltransferase genes after examining some of the available information

on yeast mutants related to phospholipid synthesis. When the CDP-DAG pathway is suppressed

by inositol in the yeast mutant ise (13), the activity of choline transporter (CTR1) becomes

essential for phospholipid biosynthesis through the CDP-choline pathway (14, 15). The ise ctr1

double mutant cannot grow on high inositol medium even in the presence of a choline

supplement (14). Such a growth defect is apparently caused by a reduced synthesis of

phosphatidylcholine (PC). It was subsequently reported that a choline transporter suppressor

gene, SCT1, corresponding to ORF YBL011w (also annotated as YBL03.09), when expressed

via a multicopy vector, could complement the cell growth defect which resulted from the

deficiency in choline transport in ise ctr1 (16). However, SCT1 cannot suppress the growth

defect of CTR1 null mutant, and over expression of SCT1 did not appear to restore choline

transport activity. It was suggested that Sct1p might stabilize the mutant form of the choline

transporter (16). However, one could also envisage Sct1p as being a positive factor involved in

the biosynthetic pathway leading to PC synthesis. This would appear to be a plausible

explanation if a small amount of choline is available, as this seems to be the case in ctr1. In the

4

by guest on Novem

ber 30, 2020http://w

ww

.jbc.org/D

ownloaded from

present study, we demonstrate that the choline transporter suppressor, Sct1p, encoded by

YBL011w, and a closely related protein encoded by YKR067w, are two yeast sn-1 acyltransferases

catalyzing both G-3-P and DHAP acylation. A point mutation in YKR067w results in a

deficiency of the acyltransferase activities in the yeast mutant TTA1. As evident in the studies of

other acyltransferases, identification of these two initial enzymes of the glycerolipid pathway is

crucial for understanding the regulation of phospholipid biosynthesis in yeast, and will hopefully

facilitate similar endeavors in other eukaryotes, including plants.

EXPERIMENTAL PROCEDURES

Yeast Strains and Culture conditions- The haploid gene disruption strains YBL011w::kanMX4

(BY4742, Matα, his3C1, leu2C0, lys2C0, ura3C0, YBL011w::kanMX4) and

YKR067w::kanMX4 (BY4742, Matα, his3C1, leu2C0, lys2C0, ura3C0, YKR067w::kanMX4),

the diploid YBL011w gene disruption strain (BY4743 Mat a/α, his3∆/ his3∆, leu2∆0/leu2∆0,

lys2∆0/LYS2, MET15/met15∆0, ura3∆0/ura3∆0, YBL011w::kanMX4/YBL011w::kanMX4),

and the wild-type strains BY4742 (Matα, his3C1, leu2C0, lys2C0, ura3C0) and DBY746 (Matα,

his3-C1, leu2-3, leu2-112, ura3-52, trp1-289) were purchased from Euroscarf. The TTA1

mutant (Matα, his3-C1, leu2-3, leu2-112, ura3-52, trp1-289) was kindly provided by Dr.

Robert M. Bell (10). Cells were cultured at 30 oC in YPD medium containing 1% Bacto-yeast

extract, 2% Bacto-peptone, and 2% glucose (Sigma).

Sequence analysis of YBL011w and YKR067w in TTA1 and DBY746- Genomic DNA (150 ng)

5

by guest on Novem

ber 30, 2020http://w

ww

.jbc.org/D

ownloaded from

from TTA1 and its parental strain DBY746 was used, respectively, to amplify the coding regions

of YBL011w and YKR067w genes. PCR amplification was performed in a 50 µl PCR reaction

containing 0.2 mM dNTPs, 0.2 µM primers, and 2.5 units pfu DNA polymerase (Stratagene).

The primers used for the amplification of YBL011w and YKR076w were 5’-

ATGCCTGCACCAAAACTCACGGAG-3’ and 5’-CTACGCATCTCCTTCTTTCCCTTC-3’,

and 5’-ATGTCTGCTCCCGCTGCCGATCAT-3’ and 5’-

TCATTCTTTCTTTTCGTGTTCTCT-3’, respectively. The PCR program employed was as

follows: initial dwell time of 2 min at 94 oC, then 32 cycles of denaturation at 94 oC for 30 s,

annealing at 60 oC for 30 s and extension at 72 oC for 3 min, followed by extension at 72 oC for

7 min. The amplified DNA fragments were cloned into pCR2.1-TOPO vector (Invitrogen)

following the addition of a single 3’ deoxyadenosine through Taq DNA polymerase treatment,

and fully sequenced using an automated DNA sequencer (Applied Biosystems 373).

Construction of GAT1 and GAT2 expression vectors - Two pairs of primers, 5’-

GGATCCAACATGTCTGCTCCCGCTGCCGATCAT-3’ and 5’-

CTCGAGTCATTCTTTCTTTTCGTGTTCTCT-3’ for GAT1 and the gat1 allele from TTA1, and 5’-

GGATCCAACATGCCTGCACCAAAACTCACGGAG-3’ and 5’-

CTCGAGCTACGCATCTCCTTCTTTCCCTTC-3’ for GAT2 gene, were designed to include BamH I and Xho I

restriction sites (underlined). The amplified DNA fragments were first cloned into vector

pCR2.1-TOPO (Invitrogen). The orientation of the insert was determined by restriction enzyme

digestion. Plasmids containing GAT1, gat1, and GAT2 were designated as GAT1/pCR2.1-

TOPO, gat1/pCR2.1-TOPO and GAT2/ pCR2.1-TOPO, respectively.

6

by guest on Novem

ber 30, 2020http://w

ww

.jbc.org/D

ownloaded from

To construct bacterial expression vectors, the coding regions of GAT1, gat1 and GAT2

were recovered by digestion of GAT1/pCR2.1-TOPO, gat1/pCR2.1-TOPO and GAT2/

pCR2.1-TOPO with BamH I. Purified DNA fragments were inserted into pQE60 (Qiagen) and

then transformed into E.coli DH5α. Prior to transforming the resulting plasmids GAT1/pQE60,

gat1/pQE60 and GAT2/ pQE60 into BB26-36 (17), correct orientation and in-frame fusion of the

inserts were confirmed by sequencing.

To construct yeast expression vectors, coding regions of GAT1 and GAT2 genes were

excised from GAT1/pCR2.1-TOPO and GAT2/ pCR2.1-TOPO through digestion with BamH I

and Xho I and inserted into vector pYES2 (Invitrogen). The integrity of the constructs, GAT1/

pYES2 and GAT2/ pYES2, was verified by sequencing. Transformation of pYES2 and the

recombinant pYES2 plasmids into gat1∆ strain was performed using lithium acetate according to

the standard protocol (18).

Expression of GAT1 and GAT2(SCT1) in E. coli BB26-36 and yeast- Single colonies

containing plasmids GAT1/pQE60, gat1/pQE60 and GAT2/ pQE60, were cultured in 2 ml LB

medium supplemented with 0.4% glucose, 0.1% glycerol and 60 µg/ml ampicillin. After

incubation at 37oC for 6 hr, the cultures were transferred to 50 ml of fresh medium, and allowed

to grow until the cell density reached OD600 =0.1. IPTG was then added to a final concentration

of 0.1 mM, and the cells were grown at 28oC for an additional 12 hr to induce protein

expression. Cells were harvested by centrifugation at 5000 xg for 5 min, washed with 50 mM

Tris-HCl (pH 8.0) and resuspended in lysis buffer (50 mM This-HCl, pH 8.0, 1 mM EDTA,

7

by guest on Novem

ber 30, 2020http://w

ww

.jbc.org/D

ownloaded from

1mM DTT, 10% glycerol). After treatment with 100 µg/ml lysozyme (Sigma) for 30 min on ice,

the suspension was sonicated six times on ice with a 15-second burst. The lysate was spun at

2000 xg for 5 min to pellet cell debris, and the supernatant was used for enzyme assays.

Expression of the proteins was confirmed through SDS-PAGE.

To over-express GAT1 and GAT2 in yeast, single colonies carrying pYES2 (plasmid-

only control) or GAT1/ pYES2 and GAT2/ pYES2 were inoculated in 10 ml SD-uracil medium

with 2% glucose. After incubation at 30oC for 30 hr, the cells were harvest by centrifugation at

1500 xg for 5 min, and resuspended in SD-uracil medium with 1% raffinose and 2% galactose

(SD induction medium). The cells were then diluted with 50 ml of SD induction medium to

obtain a cell density of OD600= 0.6. After incubation at 30oC for 7 h to induce the protein

expression, the cells were harvested by centrifugation at 1500 xg for 5 min. For preparation of

the yeast homogenates, the cell pellets were washed with 10 volumes of distilled H2O, and then

immediately frozen in liquid nitrogen and stored at -80oC until use.

Yeast homogenates were prepared with glass beads according to standard method (18). Yeast

lysate in buffer (50 mM This-HCl, pH 8.0, 1 mM EDTA, 1mM DTT, 10% glycerol) was spun at

2500 xg, 4oC for 5 min to pellet large cell debris, and the supernatant was used directly for

enzyme assays.

Protein concentration was determined using Bio-Rad Dc protein assay regents (BIO-

RAD) and bovine serum albumin as a standard.

Enzyme Activity assays- G-3-P acyltransferase activity was assayed at room temperature for 10

8

by guest on Novem

ber 30, 2020http://w

ww

.jbc.org/D

ownloaded from

min in a 200 µl reaction mixture containing 400 µM [14C] glycerol 3-phosphate (2.5 nCi/nmol),

45 µM palmitoyl-CoA, 75 mM Tris-HCl (pH 7.5), 1 mM DTT, and 2 mM MgCl2. The reaction

mixture was stopped by extracting the mixture with 3 ml of chloroform-methanol (1:2, v/v) in

the presence of 600 µl of 1% HClO4. After a repeated extraction with another 1ml chloroform

and 1 ml 1% HClO4, the lower organic phase of the Bligh and Dyer extract (19) was washed

three times with 2 ml 1% HClO4. An aliquot of the chloroform phase containing the glycerolipid

fraction was dried under nitrogen, and subjected to scintillation counting for radioactivity.

Results shown are the means ± S.E. from at least three independent assays. To confirm the

reaction products, the lipid extracts were subjected to TLC in a solvent system of chloroform /

methanol / acetic acid / 5% aqueous sodium bisulfite (100:40:12:4). The Rf values for LPA and

PA were 0.33 and 0.90, respectively (20).

DHAP acyltransferase activity was measured essentially as described by Bates and

Saggerson (21) with minor modifications. The reaction was terminated by the addition of 0.8 ml

1% HClO4, followed by extraction with 3ml chloroform-methanol (1:2, v/v) and 1 ml

chloroform. The lower phase of the Bligh and Dyer extract was washed three times with 2 ml 1%

HClO4, and the radioactivity measured through scintillation counting. The products were

separated by TLC with the above solvent systems. The Rf value of 1-acyl-DHAP in this system

was 0.20 (20).

Lipid Analysis- Yeast cell cultures at late logarithmic phase were disrupted with glass beads.

Total lipids were extracted using the procedure of Folch et al (22). Separation of the

9

by guest on Novem

ber 30, 2020http://w

ww

.jbc.org/D

ownloaded from

phospholipids was performed with two-dimensional TLC on Silica Gel-60 plates and developed

in solvent systems as described (23). Phospholipids were visualized with iodine vapor through a

nitrogen stream, scraped off the TLC plates, and transmethylated directly with methanolic-HCl.

Fatty acid methyl esters derived from each of the lipid species were analyzed and quantified by

gas chromatography. From these data the mole percentages of the analyzed phospholipids were

calculated for each lipid.

Disruption of GAT1 gene in GAT2(SCT1) null mutants- Plasmid GAT1/pCR 2.1-TOPO was

digested with Acc III and PinA I to delete a fragment corresponding to a region between nt 459

and nt 1678 in GAT1. A linker sequence with Acc III, Sma I and PinA I sites was used to link

the remaining fragments of the GAT1 gene. The yeast URA3 gene was then inserted into the

Sma I site of the linker to generate a GAT1 knockout vector. The DNA fragment containing the

whole yeast URA3 gene flanked by two partial GAT1 sequences was amplified using the vector

as a template with primers located at 5’ and 3’ ends of the GAT1 coding region (5’-

ATGTCTGCTCCCGCTGCCGATCAT-3’ and 5’-TCATTCTTTCTTTTCGTGTTCTCT-3’).

The same amount of the purified PCR product (3 µg) was used to transform the diploid

gat2(sct1)∆/gat2(sct1)∆ strain and the haploid gat2[sct1]∆ strain, respectively. More than 300

transformants were obtained from diploid knockout experiments. The transformants were

successively plated onto the uracil-dropout selective medium, and the stable transformants

bearing one disrupted GAT1 and one undisrupted GAT1 were identified by PCR and restriction

analysis of the PCR products. Several repeated experiments with the haploid gat2(sct1)∆ strain

did not result in any viable gat1∆gat2(sct1)∆ colonies. Sporulation and tetrad analysis were

performed according to standard protocol (18).

10

by guest on Novem

ber 30, 2020http://w

ww

.jbc.org/D

ownloaded from

RESULTS

Identification of Putative G-3-P Acyltransferase ORFs

Previous studies have shown that SCT1 gene is constitutively expressed in yeast. Its

encoded product is most likely a membrane protein since it was predicted to contain a cluster of

three potential membrane-spanning domains (16). Our sequence analysis revealed that Sct1p

encoded by YBL011w, and a protein encoded by YKR067w which displays 31% sequence

identity to Sct1p, contained segments with similarities to conserved domains of known

acyltransferases (24). A portion of their deduced amino acid sequences is aligned with those of

the membrane-bound G-3-P acyltransferases from Escherichia coli and the mitochondrial G-

3-P acyltransferase from Rattus norvegicus as shown in Fig. 1A. Two short segments from

YBL011wp and YKR067wp resemble the conserved motifs III and IV, respectively, of G-3-P

acyltransferases (24). The region corresponding to motif III is accentuated by a stretch of 6

amino acids (IFPEGG) that is conserved in YBL011wp and YKR067wp. The structural

similarity between the yeast proteins and other known membrane based G-3-P acyltransferases

can be further inferred by hydropathy plot predictions as shown in Fig. 1B. Based on these

analyses, we hypothesized that YBL011wp and YKR067wp are candidates for the so far elusive

membrane-bound sn-1 fatty acyltransferase.

Disruption of YBL011w and YKR067w Resulted in Reduced G-3-P Acyltransferase Activity

The above observations led us to investigate if disruption of the open reading frames

YBL011w and YKR067w would affect G-3-P acyltransferase activity in yeast. Haploid strains with

11

by guest on Novem

ber 30, 2020http://w

ww

.jbc.org/D

ownloaded from

targeted disruption in YBL011w (EUROSCARF accession no. Y13037) and YKR067W

(EUROSCARF accession no. Y15983) were acquired from the collection of deletion strains at

EUROSCARF. Neither strain displayed any abnormal growth phenotype when examined on

solid or in liquid media. Nor was there any apparent sensitivity to temperature or inositol found

upon the disruption of the respective open reading frames. Previously, it was shown that a

dramatic reduction of G-3-P acyltransferase specific activity could be easily detected even a

total homogenate of the yeast strain TTA1 was directly used for enzyme assays (12). We thus

decided to examine the effect of the gene disruptions on G-3-P acyltransferase activities

employing total yeast homogenate prepared after a brief spin at 2500 xg. As shown in Fig. 2, G-

3-P acyltransferase activity was clearly reduced in both gene disruption strains. The total G-3-P

acyltransferase activity in the YBL011w disruption strain was reduced by one third in

comparison to the parental strain. Disruption of YKR067w, on the other hand, had a more

striking effect, leaving a residual enzyme activity at about one eighth of the control (Fig. 2).

Under our assay conditions, the level of residual enzyme activity in the YKR067w gene

disruption strain is very close to the G-3-P acyltransferase activity for strain TTA1 (data not

shown), even though they are derived from somewhat different genetic backgrounds.

A Missense Mutation in YKR067W Gene Accounts for the Defect in G-3-P Acyltransferase

and DHAP Acyltransferase Activities in the TTA1 Mutant

To further investigate the implications of these initial results, we examined if the mutation

leading to the deficiencies in G-3-P acyltransferase and DHAP acyltransferase activities in the

mutant TTA1 occurred in either YBL011w or YKR067w (10). The coding regions of the two

12

by guest on Novem

ber 30, 2020http://w

ww

.jbc.org/D

ownloaded from

genes were amplified by Pfu DNA polymerase-based PCR using genomic DNA isolated from

TTA1 and its parental strain DBY746. Several nucleotide polymorphic differences were found in

both genes between the S. cerevisiae strains. The sequences of YBL011w and YKR067w from

DBY746 were deposited into the EMBL database under the accession of AJ314608 and

AJ311354, respectively. Direct sequencing of the purified PCR products and subsequent

sequencing of the PCR fragment cloned into a vector plasmid demonstrated that there was no

nucleotide sequence change in YBL011w between TTA1 and DBY746. On the other hand,

analysis of YKR067w revealed the presence of one nucleotide change from G to A at position

785 in the mutant TTA1, which is predicted to result in the substitution of aspartic acid for

glycine at amino acid position 262 of the encoded protein. Significantly, this amino acid

substitution occurred in the segment exhibiting high similarity to the conserved motif III of

acyltransferases (24). This result indicates that the deficiency of acyltransferase activity in TTA1

is attributed to this missense mutation, and thereby suggested that YKR067w encodes for a G-

3-P acyltransferase. The residual G-3-P acyltransferase activity in TTA1 is comparable to that

of the YKR067w knockout strain, suggesting that the mutation in TTA1, even in the form of a

single amino acid change, completely abolished the activity of this enzyme. This result is in

agreement with previous observations of the E. coli G-3-P acyltransferase, where a change in

the amino acid sequence in the conserved domain III from YFVEGGRSRTGR to

YFVELGRSRTGR completely eliminated the enzyme activity (24). Our results further support the

functional importance of these conserved sequence domains in fatty acyltransferases.

Previously it has been demonstrated that the defect in TTA1affected mainly the

acyltransferase activities of the lipid particle preparations. It thus can be inferred that

13

by guest on Novem

ber 30, 2020http://w

ww

.jbc.org/D

ownloaded from

YKR067wp is the lipid particle G-3-P acyltransferase. In accordance with the nomenclature

proposed by Athenstaedt and Daum (12), we named YKR067wp as Gat1p. The protein Sct1p

encoded by YBL011w, which has structural properties of a membrane protein (16), should be

localized in other cytoplasmic membrane compartments. Therefore we designated it herein as

Gat2p. The genes corresponding to YKR067w and YBL011w were named GAT1 and GAT2,

respectively.

Heterologous Expression of GAT1 and GAT2 in E. coli

Reconstitution of purified acyltransferases for enzyme activity has often proven to be

inefficient. In an attempt to establish a direct enzyme-protein relationship, GAT1 and

GAT2(SCT1) as well as the gat1 allele from TTA1 were inserted into expression vector pQE60

(Qiagen), and introduced into E. coli (plsB) strain BB26-36 (17). BB26-36 has a mutation in

plsB that gives rise to a G-3-P acyltransferase with altered properties, particularly, a lower

specific activity (25, 26). This strain had been used as a convenient host to express G-3-P

acyltransferase genes for enzyme property assessment (24). As shown in Fig. 3, G-3-P

acyltransferase activity in this plsB mutant expressing GAT1 was more than six times higher

than that of the control. In contrast, expression of gat1 in the plsB mutant showed no enzyme

activities beyond the control.

E. coli strain BB26-36 has a G-3-P auxotrophic phenotype as a result of a marked

increase in the apparent Km of the G-3-P acyltransferase for G-3-P (17, 24, 26). However,

expression of the GAT1 using both pQE60 and pET28a vectors in strain BB26-36 failed to

complement this defect. In addition, expression of the GAT2(SCT1) appeared to be extremely

14

by guest on Novem

ber 30, 2020http://w

ww

.jbc.org/D

ownloaded from

deleterious to the host cells. The growth of the cells expressing this gene was slower by a factor

of two when compared to the cells harboring the control vector. We have not been able to detect

any increase in G-3-P acyltransferase activity through expression of GAT2(SCT1) in E. coli.

Substrate Specificity of the Gat1p and Gat2p(Sct1p): G-3-P/DHAP Dual Substrate Specific

Acyltransferase

Due to the apparent difficulties involved in the reconstitution of enzyme activities of

membrane-bound acyltransferases, we decided to adapt a strategy based on the low G-3-P

acyltransferase background of the gat1∆ strain to investigate the substrate specificities of Gat1p

and Gat2p(Sct1p) with respect to G-3-P and DHAP. The two genes were expressed using a

multiple copy vector pYES2 under the control of GAL+ promoter. Specific activities of G-3-P

and DHAP acyltransferase of the two proteins were evaluated using palmitoyl CoA as the fatty

acyl donor (Fig. 4). Over-expression of GAT1 resulted in a net increase of 4.4 and 3.2

nmol.min-1 .mg protein–1 in G-3-P acyltransferase and DHAP acyltransferase activities, respectively.

Likewise, significant increases in G-3-P acyltransferase and DHAP acyltransferase activities

were also evident in the gat1∆ strain over-expressing GAT2(SCT1) (Fig. 4). The observed

increases in the specific activities of G-3-P and DHAP acyltransferases indicate that Gat1p and

Gat2p(Sct1p) can efficiently utilize both G-3-P and DHAP as substrates, thereby providing

direct evidence that the two yeast sn-1 acyltransferses are G-3-P/DHAP dual substrate

acyltransferases. It is also consistent with the view of Athenstaedt and Daum (27) that Ayr1P, a

major component of lipid particles which functions as a 1-acyl-DHAP reductase (which results

in production of LPA), works coordinately with Gat1p to carry out the DHAP-dependent

15

by guest on Novem

ber 30, 2020http://w

ww

.jbc.org/D

ownloaded from

glycerolipid pathway in yeast lipid particles. As shown in Fig. 4, Gat1p displayed almost the

same level of specific enzyme activities with regard to G-3-P and DHAP, while Gat2p(Sct1p)

clearly preferred G-3-P even though DHAP was also an effective substrate.

Substrate Specificity of the Gat1p and Gat2p(Sct1p): Fatty Acyl Preferences

Fatty acid substrate specificity of acyltransferases plays an important role in determining

sterospecific distributions of fatty acyl groups in glycerolipids. Substrate preference in relation to

saturated and unsaturated fatty acids has also been frequently implicated in regulation of

temperature-dependent incorporation of fatty acids into phopholipids. To investigate the fatty

acyl substrate preferences of Gat1p and Gat2p(Sct1p), specific activities towards palmitoyl-

CoA, palmitoleoyl-CoA, stearoyl-CoA, and oleoyl-CoA were compared using the gat1∆ strain

expressing GAT1 and GAT2(SCT1), respectively. As shown in Fig. 5, Gat1p could efficiently

utilize all four fatty acyl substrates, with a noticeably lower specific activity towards 18:0-CoA.

In general, the characteristics of the fatty acyl specificity of Gat1p are similar to that of the

mammalian microsomal G-3-P acyltransferase, which is also capable of utilizing a broad range

of acyl-CoAs (8). In contrast, Gat2p(Sct1p) exhibited marked preference for 16 carbon fatty

acids. Moreover, both enzymes appeared to prefer unsaturated fatty acids over saturated ones.

Phospholipid and Fatty Acid Profiles of gat1∆ and gat2(sct1)∆ Strains

To investigate the respective roles and relative contributions of the two acyltransferases

to phospholipid metabolism, the steady-state levels of phospholipids from the gat1∆ and

gat2(sct1)∆ strains were compared with those of the parental strain. In accordance with data

16

by guest on Novem

ber 30, 2020http://w

ww

.jbc.org/D

ownloaded from

reported for the TTA1 strain (12), the size of the phosphatidic acid (PA) pool in the gat1∆ strain,

measured as molar percent of total phospholipids, was reduced to less than half of that of the

parental strain (Fig 6). Similar reduction of the PA pool in the gat2(sct1)∆ strain was also

observed. There is also a detectable change in the relative abundance of PS and PI, with a PS/PI

molar ratio elevated from 0.33 in the parental strain to 0.60 and 0.55, respectively, in the gat1∆

and gat2(sct1)∆ strains, respectively.

In light of the fatty acyl substrate specificities of the two acyltransferases, we also

examined the fatty acid compositions of the major phospholipid species in gat1∆ and

gat2(sct1)∆. The data presented in Table 1 can be summarized as follows: (i) lack of Gat1p in yeast did not

seem to have a significant effect on the total fatty acid profiles of PC, PS and PI. However, a

decrease in 16:1 fatty acid was observed in PE, and the reduction is proportionally compensated

by increases in both 16:0 and 18:1; (ii) the absence of Gat2p(Sct1p) impacted fatty acid

compositions in all four major phospholipid species. In general, the gat2(sct1)∆ mutant had

proportionally less 16:0, and such a decrease in 16:0 was offset by increases in the proportions of

other fatty acids, particularly 18:0.

Null mutations of GAT1 and GAT2(SCT1) are synthetically lethal to yeast

To investigate the growth and biochemical phenotype of the gat1∆ gat2(sct1)∆ double

null mutant, deletion of GAT1 was carried out in a diploid gat2(sct1)∆/gat2(sct1)∆ strain.

Subsequent tetrad analysis showed that only two colony-forming spores were generated from

each tetrad, and all of the viable colonies carried the intact GAT1. When GAT1 deletion was

attempted with the haploid gat2(sct1)∆ strain, repeated experiments failed to generate any viable

17

by guest on Novem

ber 30, 2020http://w

ww

.jbc.org/D

ownloaded from

double mutant. Taken together, it indicates that gat1∆ and gat2(sct1)∆ are synthetically lethal to

yeast.

Discussion

The importance of LPA formation as the committed and potentially rate-limiting step in

phospholipid synthesis has been well recognized. To obtain a better understanding of the initial

acylation reaction of the glycerolipid pathway in yeast and other eukaryotes, we carried out the

identification of G-3-P acyltransferase genes. Our rationale was based on previous studies of the

yeast mutant ise, a mutant involved in the regulation of the phosphatidylethanolamine

methylation pathway (13), and the choline transporter mutant ctr1 (14). Our molecular and

biochemical studies demonstrated that YBL011w and YKR067w, designated in this work as

GAT2 and GAT1, respectively, are genes encoding for cytoplasmic membrane-bound G-3-P

acyltransferases.

The ise mutant is a conditional choline auxotrophic mutant. Its growth is inhibited by high

concentrations of inositol, but this defect can be suppressed by supplying the inositol-containing

medium with choline (13). The growth defect of ise mutant in response to inositol has been

shown to be due to a dramatic decrease in the phosphatidylethanolamine (PE) methyltransferase

activity (13). Choline supplementation suppresses the growth defect, but cannot reverse the

decrease in the enzyme activities of PE methyltransferases imposed by high amounts of inositol

(13), indicating that the supply of choline may lead to an increase in PC synthesis via the CDP-

choline pathway. A choline transporter mutant, ctr1, had a marked decrease in choline supply,

and thereby a weakened CDP-choline pathway for PC synthesis. The ise ctr1 double mutant

18

by guest on Novem

ber 30, 2020http://w

ww

.jbc.org/D

ownloaded from

showed a growth defect when high levels of inositol is present in the medium, even in the

presence of choline (14). Taken together, it strongly suggests that the combination of a crippled

CDP-choline pathway and a PE methylation pathway is the root cause for the growth defect of

ise ctr1 at high levels of inositol.

Previously, GAT2 (SCT1) was found to suppress the growth defect of the ise ctr1 mutant in

the presence of inositol and choline (16). However, the precise mechanism by which

Gat2p(Sct1p) suppresses the ise and ctr1 mutations had not been defined. Since Gat2p(Sct1p)

exhibits properties of neither a choline transporter (16) nor a methyltransferase, we reasoned that

it might exert its role through enhancing certain steps in the phospholipid pathways. This is

consistent with the notion that ctr1 apparently had some residual level of choline uptake (15),

and that the suppression effect of Gat2p(Sct1p) was in a choline-dependent manner (16).

Moreover, GAT2(SCT1) can only exert its effect via a multicopy vector, suggesting that it

functions most like an enzyme rather than a regulatory protein in stimulating the CDP-choline

pathway of PC synthesis (16).

We speculated that Gat2p(Sct1p) and the closely related Gat1p, encoded by YKR067w,

were G-3-P acyltransferases, in part based on the analysis of their sequences in which we

uncovered two regions similar to the conserved motifs of known acyltransferases (24). Indeed,

disruption of either GAT1 or GAT2(SCT1) resulted in a reduction in the total cellular G-3-P

acyltransferase activities. Our assumption was further substantiated by a point mutation revealed

in gat1 of the G-3-P acyltransferase mutant TTA1. In addition, over-expression of the GAT1

and GAT2(SCT1) genes in the gat1∆ strain, which has a low G-3-P acyltransferase background,

led to highly elevated enzyme activities. Finally, expression of GAT1 in E. coli strain BB26-36

19

by guest on Novem

ber 30, 2020http://w

ww

.jbc.org/D

ownloaded from

demonstrated a direct enzyme-protein relationship. Although our experiments in expression of

GAT2(SCT1) in E. coli failed to indicate any increase in G-3-P acyltransferase activity, this is not

surprising in a heterologous system for a membrane protein. A major structural difference

between Gat1p and Gat2p(Sct1p) lies at the C-terminals (16). We suspect that either the poly-

glutamic acid track or the so called PEST-like region, which is rich in proline (P), glutamic acid

(E), serine (S), and threonine (T), and is often involved in fast-turnover of proteins (16), may

contribute to the particular difficulties in the reconstitution of enzyme activity in E. coli.

However, since both knockout and over expression of GAT2(SCT1) in yeast generated results

with regard to G-3-P acyltransferase activity resembling that of those with GAT1, we believe it

leaves little doubt that Gat2p(Sct1p) is a second isoform of the enzyme in yeast. It should be

noted that even though in vitro assays detected a several-fold increase in enzyme activity in the

GAT1-expressing E. coli lysate, it is unlikely that the protein is at a correct configuration in

membranes in vivo, since it failed to complement the G-3-P auxotrophic phenotype of BB26-

36. The sonication procedure used in preparing the bacterial lysate might somewhat assist in the

reconstitution of enzyme activity for Gat1p by forming micelles.

Studies of Tillman and Bell found that the mutation in TTA1 led to deficiencies in both

G-3-P and DHAP acyltransferase activities (10). Using highly purified lipid particles of TTA1,

Athenstaedt and Gaum also reached the same conclusion that G-3-P and DHAP acyltransferase

activities are attributable to a single protein (12). Our present study provides direct evidence to

demonstrate that both of these two sn-1 acyltransferases in yeast catalyze G-3-P and DHAP

acylation.

As noted before in studies with TTA1 (12), the PA pool is reduced in both gat1∆ and

20

by guest on Novem

ber 30, 2020http://w

ww

.jbc.org/D

ownloaded from

gat2(sct1)∆ strains. Even though disruption of neither gene seemed to result in a significant

reduction of the steady-state level of PC, an increase in PS/PI ratio was consistently observed.

The alterations of fatty acid composition in phospholipids resulting from the lack of either of the

two isforms are generally consistent with their respective fatty acyl preferences. Gat1p, which

does not exhibit particular substrate preference, has limited influence on fatty acid profiles.

Gat2p(Sct1p), on the other hand, due to its apparent preference towards 16-carbon saturated

fatty acids, plays a major role in gauging the fatty acid composition of phospholipids.

The conclusion of Gat1p being the lipid particle G-3-P acyltransferase was discerned

from the molecular defect revealed in YKR067w in TTA1. The substrate specificity of Gat1p

with regard to G-3-P / DHAP is also in line with the data of the lipid particle acyltransferase

reported from other biochemical studies. Studies of Athenstaedt et al (12) have shown that in

addition to the lipid particle acyltransferase, redundant systems of phospholipid biosynthesis are

present in both the microsomal fraction and in mitochondrial fractions. The mitochondrial sn-1

acyltransferase displays a low ratio of G-3-P acyltransferase to DHAP acyltransferase activity,

and represents a distinct DHAP acyltransferase (12). Since Gat2p(Sct1p) clearly prefers G-3-P

to DHAP, it should be present in the endoplasmic reticulum. The fact that Gat2p(Sct1p)

functioned as a choline transporter suppressor also suggests that it is more likely to be localized

in plasma or cytoplasmic membrane systems. The synthetic lethal phenotype of gat1∆gat2(sct1)∆

suggests that Gat1p and Gat2p(Sct1p) are the major, if not the only, sn-1 acytransferases

existing in the cytosolic compartment. This study has provided molecular information for future

efforts to address the complexity of phospholipid biosynthesis in eukaryotic cells.

21

by guest on Novem

ber 30, 2020http://w

ww

.jbc.org/D

ownloaded from

Acknowledgements:

We thank Yu Fu and Dr. Wei Xiao for technical assistance on tetrad analysis. We are

grateful for helpful discussions with Drs. R. Datla, V. Katavic, P. Covello and D. Taylor and

their critical readings of the manuscript. This research was supported by the National Research

Council–Plant Biotechnology Institute core program. This is NRCC publication No.43807.

References

1. Daum, G., Lees, N.D., Bard, M., and Dickson, R. (1998) Yeast 14, 1471-1510

2. Carman, G. M., and Henry, S. A. (1999) Prog. Lipid Res. 38, 361-399

3. Moolenaar, W.H. (1995) J. Biol. Chem. 270,12949-12952

4. English, D., Cui, Y., and Siddiqui, R. (1996) Chem Phys Lipids 80, 117-132

5. Dircks, L. and Sul, H. S. (1999). Prog. Lipid Res. 38, 461-479

6. Christiansen, K (1978) Biochim. Biophys. Acta. 530(1), 78-90

7. Wilkison, W. O., and Bell, R. M. (1997) Biochim. Biophys. Acta. 1348, 3-9

8. Dircks, L. K., and Sul, H. S. (1997) Biochim. Biophys. Acta. 1348, 17-26

9. Murata, N., and Tasaka, Y. (1997) Biochim. Biophys. Acta. 1348, 10-16

10. Tillman, T. S., and Bell, R. M. (1986) J. Biol. Chem. 261(20), 9144-9

11. Athenstaedt, K., and Daum, G. (1997) J. Bacteriol. 179(24), 7611-6

12. Athenstaedt, K., Weys, S., Paltauf, F., and Daum, G. (1999) J. Bacteriol. 181(5), 1458-

63

13. Yamashita, S., and Oshima, A. (1980) Eur. J. Biochem. 104, 611-616

14. Nikawa, J., Tsukagoshi, Y., and Yamashita, S. (1986) J. Bacteriol. 166, 328-330

22

by guest on Novem

ber 30, 2020http://w

ww

.jbc.org/D

ownloaded from

15. Nikawa, J., Hosaka, K., Tsukagoshi, Y., and Yamashita, S. (1990) J. Biol. Chem. 265, 15996-

16003

16. Matsushita, M., and Nikawa, J. (1995) J. Biochem. 117, 447-451

17. Bell, R. M. (1974) J. Bacteriol. 117, 1065-1076

18. Ausubel, F. M., Brent, R., Kingston, R. E., Moore, D. D., Seidman, J. G., Smith, J. A.,

and Struhl, K. (1994) Current Protocols in Molecular Biology. John Wiley & Sons, Inc.

p13.0.1-13.13.9

19. Bligh, E. G., and Dyer, W. J. (1959) Can. J. Biochem. Physiol. 37, 911-917

20. Hajra, A. K., and Burke, C. (1978) J. Neurochem. 31, 125-134

21. Bates, E. J., and Saggerson, E. D. (1979) Biochem. J. 182, 751-762

22. Folch, J. M., Lees, M. and Sloane-Stanley, G. H. (1957) J. Biol. Chem. 226, 497-509

23. Morash, S. C., MacMaster, C. R., Hjelmstad, R. H., and Bell, R. M. (1994) J. Biol. Chem. 269,

28769-28776

24. Lewin, T. M., Wang, P., and Coleman, R. A. (1999) Biochem. J. 38, 7564-5771

25. Lightner, V. A., Larson, T. J., Tailleur, P., Kantor, G. D., Raetz, C. R. H., Bell, R. M., and Modrich, P.

(1980) J. Biol. Chem. 255, 9413-9420

26. Heath, R. J., and Rock, C. O. (1999) J. Bacteriol. 181, 1944-1946

27. Athenstaedt, K., and Daum, G. (2000) J. Biol. Chem. 275, 235-240

28. Shen, H., and Dowhan, W. (1996) J. Biol. Chem. 272, 11215-11220.

23

by guest on Novem

ber 30, 2020http://w

ww

.jbc.org/D

ownloaded from

Figure legend

Fig. 1. Conserved motifs of YBL011w and YKR067w encoding proteins in comparison to

known glycerol-3-phosphate acyltransferase sequences. (A) Alignment of YBL011wp

and YKR067wp with partial sequences of G-3-P acyltransferase from Escherichia coli

(PlsB; accession no. P00482) and Rattus norvegicus (RGPAT; accession no.

NP_058970), using MegAlign program from the software package DNAstar. Identical

amino acid residues are highlighted in shade. The glycine residue in Ykr067wp, which is

converted to an aspartic acid as a result of a point mutation in TTA1, is marked with an

asterisk. (B) Hydropathy profiles of the acyltransferases predicted with the Kyte-

Doolittle algorithm. An average of 9 residues is plotted for hydropathy value.

Hydrophilic regions are defined as positive values, and hydrophobic regions as negative

values. The abscissa is the residue number at the center of each stretch.

Fig. 2. G-3-P acyltransferase activity in strain BY4742 (WT), YKR067w and YBL011w gene

disruption strains. Cells of gene disruption strains YKR067w::kanMX4 (A) and

YBL011w::kanMX4 (B), as well as the parental strain BY4742 (C) were grown in YPD

medium to a late logarithmic phase and used for total homogenate preparations as

described under EXPERIMENTAL PROCEDURES. G-3-P acyltransferase (GAT)

24

by guest on Novem

ber 30, 2020http://w

ww

.jbc.org/D

ownloaded from

activities were assayed with 400 µM [14C] glycerol 3-phosphate (2.5 nCi/nmol), 45 µM

palmitoyl-CoA, 75 mM Tris-HCl (pH 7.5), 1 mM DTT, and 2 mM MgCl2 for 10 min at

room temperature. After extraction of the phospholipid products, the radioactivity was

measured.

Fig. 3. Functional expression of the wild-type (GAT1) and mutant forms (gat1) of YKR067w

gene in E. coli strain BB26-36. BB26-36 cells harboring GAT1 and gat1 expression

vectors were cultured at 37 oC in LB medium supplemented with 0.1 % glycerol, 0.4%

glucose, and 60 µg / L ampicillin. After the cell density reached OD600= 0.1, IPTG was

added to a final concentration of 0.1 mM, and the cells were grown at 28 oC for 12 hr to

induce protein expression. Cell lysates were prepared as described in EXPERIMENTAL

PROCEDURES. G-3-P acyltransferase activity was assayed with 400 µM [14C]

glycerol 3-phosphate using palmitoyl-CoA as fatty acyl donor. Background enzyme

activity in the cells bearing the control vector pQE60 was also shown.

Fig. 4. G-3-P and DHAP acyltransferase activities in the gat1∆ strain over-expressing GAT1

and GAT2(SCT1) genes. Expression vector pYES2 harboring GAT1 or GAT2(SCT1),

respectively, was introduced into the gat1∆ yeast strain, and G-3-P acyltransferase

(GAT) and DHAP acyltransferase (DHAPAT) activities were assayed upon galactose

induction. The gat1∆ strain bearing vector pYES2 was used as a control. Palmitoyl-CoA

(65µM) was used as fatty acyl donor in both enzyme assays. The concentrations of

25

by guest on Novem

ber 30, 2020http://w

ww

.jbc.org/D

ownloaded from

glycerol-3-phosphate and dihydroxyacetone phosphate used in the assays were 400 µM

and 450 µM, respectively. The enzyme reactions were carried out at 30 oC for 10 min.

Fig. 5. Fatty acyl substrate specificity of the Gat1p and Gat2p(Sct1p). GAT1 and GAT2(SCT1)

were inserted into yeast expression vector pYES2, and expressed in the gat1∆ strain. G-

3-P acyltransferase activity from cells containing vector alone was used as a control.

Fatty acyl substrates used in the assays were palmitoyl -CoA (16:0-CoA), palmitoleoyl-

CoA (16:1-CoA), stearoyl-CoA (18:0-CoA), and oleoyl-CoA (18:1-CoA).

Fig. 6. Relative phospholipid compositions of GAT1 and GAT2(SCT1) null mutants and the

wild-type strain BY4742. Wild type, gat1delta and gat2(sct1)delta strains grown in YPD

medium to a late logarithmic phase were used for lipid extraction. Separation and

quantification of phospholipids were performed as described under “EXPERIMENTAL

PROCEDURES”. The abbreviations used are: PC, phosphatidylcholine; PE,

phosphatidylethanolamine; PS, phosphatidylserine; PI, phosphatidylinositol; PA,

phosphatidic acid, DMPE, dimethylphosphatidylethanolamine.

26

by guest on Novem

ber 30, 2020http://w

ww

.jbc.org/D

ownloaded from

Table 1. Fatty acid compositions of different phospholipids in wild-type and mutant strains. BY4742: wild type parental yeast strain; YKR067w:kanMX4: the gat1∆ strain; YBL011w:kanMX4: the gat2(sct1)∆ strain. Cells grown in YPD medium to a late logarithmic phase were used for fatty acid profile analysis.

proportion of fatty acids (Mol %) Phospho- Strain lipid 16:0 16:1 18:0 18:1 BY4742 (WT) 16.06 64.17 2.76 17.0 PC YKR067w:kanMX4 17.07 63.90 2.70 16.37 YBL011w:kanMX4 10.82 66.15 4.44 18.59 BY4742 (WT) 43.93 19.07 12.37 24.62 PI YKR067w:kanMX4 43.69 18.83 11.86 25.62 YBL011w:kanMX4 36.36 19.96 15.07 28.61

BY4742 (WT) 37.90 25.73 NDa 36.36 PS YKR067w:kanMX4 38.10 26.46 ND 35.41 YBL011w:kanMX4 35.65 29.84 ND 34.51 BY4742 (WT) 20.12 48.17 ND 31.71 PE YKR067w:kanMX4 25.18 37.0 ND 37.83 YBL011w:kanMX4 18.95 47.61 2.05 31.40

a ND, not detectable.

1

by guest on Novem

ber 30, 2020http://w

ww

.jbc.org/D

ownloaded from

by guest on Novem

ber 30, 2020http://w

ww

.jbc.org/D

ownloaded from

by guest on November 30, 2020 http://www.jbc.org/ Downloaded from

by guest on November 30, 2020 http://www.jbc.org/ Downloaded from

by guest on November 30, 2020 http://www.jbc.org/ Downloaded from

by guest on November 30, 2020 http://www.jbc.org/ Downloaded from

by guest on November 30, 2020 http://www.jbc.org/ Downloaded from

Zhifu Zheng and Jitao Zoucerevisiae

dihydroxyacetone phosphate dual substrate acyltransferases in Saccharomyces The initial step of the glycerolipid pathway: Identification of glycerol-3-phosphate /

published online September 5, 2001J. Biol. Chem. 

  10.1074/jbc.M104749200Access the most updated version of this article at doi:

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

by guest on Novem

ber 30, 2020http://w

ww

.jbc.org/D

ownloaded from