Biochemical characterization of the PHARC-associated serine hydrolase ABHD12 reveals its preference for very-long-chain lipids

JBC Papers in Press. Published on September 20, 2018 as Manuscript RA118.005640

The latest version is at http://www.jbc.org/cgi/doi/10.1074/jbc.RA118.005640

Biochemical characterization of the PHARC associated serine hydrolase

ABHD12 reveals its preference for very long chain lipids

#

,#

1,

2

1,#

1,*

1

Alaumy Joshi , Minhaj Shaikh , Shubham Singh , Abinaya Rajendran , Amol

1

Mhetre , Siddhesh S. Kamat

From the ¹Department of Biology and ²Department of Chemistry, Indian Institute of Science

Education and Research (IISER) Pune, Dr. Homi Bhabha Road, Pashan, Pune 411008, India

Running title: ABHD12 prefers very long chain lipids

*To whom correspondence should be addressed: Siddhesh S. Kamat, Department of Biology,

Indian Institute of Science Education and Research (IISER) Pune, Dr. Homi Bhabha Road,

Pashan, Pune 411008, India; siddhesh@iiserpune.ac.in; Tel. +91-20-25908433.

#

Authors contributed equally

Keywords: ABHD12, PHARC, neurodegenerative disease, enzyme kinetics, Michaelis-Menten, lipase,

subcellular fractionation, very long chain lipids, hydrolase.

ABSTRACT

biochemical explanation for why very long chain

Polyneuropathy, hearing loss, ataxia, retinitis

lipids (such as lysophosphatidylserine lipids)

pigmentosa, and cataract (PHARC) is a rare

accumulate in the brains of ABHD12 knockout

genetic human neurological disorder caused by

null mutations to the Abhd12 gene, which encodes

mice, which is a murine model of PHARC.

the integral membrane serine hydrolase enzyme

ABHD12. While the role that ABHD12 plays in

Polyneuropathy, hearing loss, ataxia, retinitis

pigmentosa, and cataract (PHARC) (OMIM #

PHARC is understood, the thorough biochemical

characterization of ABHD12 is lacking. Here, we

report the facile synthesis of mono-1-(fatty)acyl-

glycerol lipids of varying chain lengths and

unsaturation, and use this lipid substrate library, to

612674) is

a

rare autosomally recessive

neurological disorder caused by homozygous or

compound heterozygous mutations in the Abhd12

gene on chromosome 20p11 in humans (1-4). This

gene encodes an integral membrane enzyme α/β

hydrolase domain containing protein # 12

(ABHD12) that belongs to the metabolic serine

hydrolase family of enzymes (5,6). The major

symptoms of PHARC include polymodal sensory

and motor defects linked to peripheral neuropathy,

hearing loss and early onset of cataract and

eventual blindness (1,2,4). Other clinical

symptoms include massive cerebellar atrophy and

demyelination of sensorimotor neurons (1,2). The

symptoms of PHARC appear in late childhood or

biochemically

characterize

recombinant

mammalian ABHD12. The substrate profiling

study for ABHD12 suggested that this enzyme

requires glycosylation for optimal activity, and

that is has a strong preference for very long chain

lipid substrates. We further validated this

substrate profile against brain membrane lysates

generated from wild type and ABHD12 knockout

mice.

Finally,

using

cellular

organelle

fractionation and immunofluorescence assays, we

show that mammalian ABHD12 is enriched on the

endoplasmic reticulum membrane, where most of

the very long chain fatty acids are biosynthesized

in cells. Taken together, our findings provide a

1

ABHD12 prefers very long chain lipids

early teenage years, worsen progressively with

age, and have no bias to gender or race (1,2). To

date, several mutations have been mapped in

human PHARC subjects, all predicted to cause

loss of ABHD12 activity and hence its biological

function (1-4,7-9). Recently, the murine model of

PHARC, i.e. the ABHD12 knockout mouse line

was generated, and extensively characterized (5).

These mice exhibited age-dependent PHARC-like

phenotypes, which included auditory and motor

deficits (5). These mice displayed heightened

neuroinflammation and cerebellar microgliosis at

time points preceding the sensorimotor defects

(5). Untargeted lipidomics on the brains of these

MAG lipase given its ability to robustly hydrolyze

the endocannabinoid 2-arachidonoyl-glycerol (2-

AG) (10). Since then, another study shows that

mammalian ABHD12 can use both 2-AG and 1-

arachidonoyl-glycerol (1-AG) at comparable

rates, with a slight preference for 1-AG as an in

vitro substrate (12). The same study also

describes, that mammalian ABHD12 can use

other long lipid chain containing mono-1-

(fatty)acyl-glycerol (1-MAG) and mono-2-

(fatty)acyl-glycerol (2-MAG) lipid substrates at

comparable enzymatic rates (12). All in vitro

substrate profiling studies taken together, show

that mammalian ABHD12 does not use

phospholipids, diacylglycerols or triacylglycerols

lipids as substrates, thus limiting the substrate

scope of this enzyme to only lyso-PS, 2-MAG and

1-MAG lipids (Figure 1) (11,12).

mice,

showed

elevated

levels

of

lysophosphatidylserine (lyso-PS) class of lipids,

suggesting that deregulated lyso-PS signaling as a

likely causative factor for the PHARC-like

symptoms displayed by these mice. Of note, very

long chain (VLC, ≥ C22) lyso-PS lipids were

massively accumulated in the brains of these mice

(5). These experiments thus suggest that lyso-PS

lipids are possibly the in vivo substrates of

ABHD12.

While these pioneering biochemical

studies describe the substrate scope of mammalian

ABHD12, it is important to note that all these

substrate-profiling studies were done at a single

substrate concentration (25 or 100 µM), and

rigorous Michaelis-Menten type enzyme kinetic

studies on mammalian ABHD12 are lacking for

any substrate except 2-AG to the best of our

knowledge (12). Also, these single concentration

substrate-profiling studies were done for only for

medium (C8 – C12) and long chain (C14 – C20)

fatty acid MAG substrates, and VLC (≥ C22) fatty

acid containing MAG substrates have not been

tested against any mammalian ABHD12 to the

best of our knowledge. This again, is because

VLC containing MAG lipids are not easily

available from commercial sources, and/or are

very expensive for such studies.

While ABHD12 prefers lyso-PS lipids as

substrates in vivo, virtually no studies are

available that describe the thorough biochemical

characterization of ABHD12 using lyso-PS lipids

as substrates. This is mostly because only a

handful of lyso-PS lipids are available

commercially, and even these are very expensive

to perform such studies. Additionally, there are

currently no facile synthetic routes to generate

lyso-PS lipids, thereby limiting their use as

substrates for biochemical characterization of this

enzyme. However, ABHD12 hydrolyzes other

lipids in vitro, making them suitable surrogate

substrates for such biochemical substrate profiling

studies. Previous studies have shown that

ABHD12 can robustly hydrolyze mono-

(fatty)acyl-glycerol (MAG) substrates in vitro

(10-12). In fact, ABHD12 was first annotated as a

Given the lack of this, we decided to

perform a rigorous substrate profiling study of

mammalian ABHD12. Towards this, here, we

describe a facile synthetic route to generate 1-

MAG lipids substrates of varying chain lengths

and unsaturation. We assay these 1-MAG lipid

2

ABHD12 prefers very long chain lipids

substrates against recombinant (human) and

endogenous (mouse brain) mammalian ABHD12,

and generate a broad substrate profile for

mammalian ABHD12. Lastly, we are the first to

report the cellular localization of ABHD12 using

biochemical cellular organelle fraction and

immunofluorescence studies, and taken together

our studies posit a biochemical explanation for the

VLC lipid (lyso-PS) accumulation seen in the

brains of ABHD12 knockout mice, the murine

model of PHARC (5).

Since ABHD12 is an integral membrane

enzyme, and no purification method is described

to date for hABHD12, we decided to perform all

assays against membrane lysates derived from

mock or hABHD12 transfected HEK293T cells

using a previously described transfection protocol

(15). We also generated the “catalytically dead”

active site mutant (S246A) for hABHD12, to use

as an additional control to mock transfected

lysates for these substrate assays (5). We

confirmed hABHD12 activity in the membrane

lysates obtained from wild type (WT) hABHD12

transfected HEK293T cells, and the lack there of

in the membrane lysates obtained from mock or

the S246A hABHD12 transfected cells, by gel

based activity based protein profiling (ABPP)

(16,17) using the fluorophosphonate rhodamine

(FP-rhodamine) probe (6,16) (Figure 3A). We

also confirmed by western blot analysis, the near

equal expression of WT and S246A hABHD12 in

their respective membrane lysates, and lack of

hABHD12 expression in mock membrane lysates

(Figure 3A). Previous studies have shown that

mammalian ABHD12 is a highly glycosylated

enzyme (10,15), and we wanted to assess whether

the glycosylation had any effect on the activity of

hABHD12. Towards this, we first assessed

whether WT hABHD12 expressed by us was

glycosylated. We pretreated HEK293T membrane

lysates overexpressing WT hABHD12 with FP-

rhodamine, and denatured these lysates by

detergent treatment and heating, and the denatured

lysates were treated with PNGaseF as per

manufacturers instructions. The resulting lysates

were visualized by gel based ABPP, and western

blot analysis for deglycosylation. We found from

both these assays that overexpressed WT

hABHD12 was indeed glycosylated, as evident

from the shift of untreated band to a lower

molecular weight band in PNGaseF treated

samples (Figure 3B). Next, we wanted to assess

whether glycosylation of hABHD12 had any

Results

Synthesis of 1-MAG lipid library

To assess the 1-MAG substrate preference

of mammalian ABHD12, we decided to

synthesize a library of 1-MAG lipids of varying

fatty acid chain lengths, and differing extent of

unsaturation. We wanted the synthesis route to be

relatively easy, possible at a small scale (low

milligram) using commercially available free fatty

acids, and able to generate good reaction yields.

Towards this, we adapted a two-step synthetic

scheme from known and well-reported synthetic

reactions (Figure 2). In the first step, using 1-

ethyl-3-(3-dimethylaminopropyl)carbodiimide

(EDC) as a coupling reagent, esterification of a

free fatty acid (from C10 to C24) with 1,2-

isopropylideneglycerol was performed (13). The

second step involved the deprotection of the

isopropylidene group using the Amberlyst-15

catalyst to yield the corresponding 1-MAG lipid

of interest (14). All reactions were performed on a

10 or 20 mg scale depending on the cost and

availability of the starting free fatty acid, and

these reactions afforded yields from 50 – 94%

(Figure 2). The complete synthesis description

and detailed compound characterization details

are available in the Supplementary Information.

Effect of glycosylation on the activity of human

ABHD12 (hABHD12)

3

ABHD12 prefers very long chain lipids

effect on its activity. We however could not assess

this by gel based ABPP, as the PNGaseF

treatment is detergent dependent, and addition of

detergent to the gel based ABPP assay protocol

prevented any active FP-rhodamine labeling of

WT hABHD12 in our experiments. We thus tested

this in an established LC-MS lipase using C18:1

lyso-PS and C18:1 1-MAG lipid substrates

(10,15). We found from this assay, that

deglycosylation of WT hABHD12 results in

complete loss of its lipase activity against both

C18:1 lyso-PS and C18:1 1-MAG, suggesting that

glycosylation of hABHD12 is critical for its lipase

activity (Figure 3C). Since we observed that

detergent treatment from this protocol causes

some loss of ABHD12 activity (~25%), we

avoided detergent in all subsequent ABHD12

activity assays.

C18:1 2-MAG as substrates (all 100 µM) and

assayed them against 20 µg of WT hABHD12

transfected HEK293T membrane lysate, and

measured lipase activity for these substrates as a

function of time. We found that there was a nice

linear relationship between the enzymatic rate and

time of the assay up to 1 hour, allowing initial

velocity measurements for enzyme kinetics

studies up to 1 hour of this lipase assay (Figure

4B). We chose 30 mins for all subsequent lipase

assays, as this allowed us enough time for any

downstream sample processing, and running

multiple reactions at once. Since we were

planning to assay different fatty acid containing 1-

MAG lipids, we wanted to confirm that there was

no difference in quantitatively measuring these

free fatty acids in our LC-MS method. Indeed this

was the case, where all free fatty acids from C10 –

C24 behaved similarly, and had a linear dynamic

range from 1 pmol to 1 nmol in our LC-MS

method (Supplementary Table 1).

Substrate profiling study against recombinant

hABHD12

Having synthesized 16 1-MAG lipid

Having established the appropriate assay

substrates of varying lipid chain lengths and

different unsaturation(s), we decided to perform

enzyme kinetics studies on recombinant

hABHD12 against this lipid substrate library,

since such study is lacking for any mammalian

ABHD12. Towards this, we first wanted to assess

the relationship between enzyme concentrations

and enzymatic rate. In this study we used C18:1

lyso-PS, C18:1 1-MAG, and C18:1 2-MAG as

substrates (all 100 µM) and assayed them against

conditions for the enzyme kinetic studies, we

tested both mock and WT hABHD12 against the

1-MAG substrate library at varying substrate

concentrations (0 – 400 µM). The corrected

enzymatic rate for WT hABHD12 for a particular

concentration for a lipid was obtained by

subtracting the corresponding mock rate at that

concentration for that lipid, and the corrected WT

hABHD12 enzymatic rates at different substrate

concentrations for a particular lipid were plotted

and fit to a classical Michaelis-Menten kinetics

equation. The kinetic constants from these

enzymatic assays for WT hABHD12 are reported

in Table 1. Based on the enzyme kinetics data

from the 1-MAG substrate profiling, in vitro,

ABHD12 has a strong preference for VLC

containing 1-MAG lipids from all the kinetic

constants i.e. V_max, K_mand V_max/K_m(Table 1).

Comparing the kinetic constants for the saturated

fatty acid 1-MAG lipids, we find a C24:0 ≈ C22:0

varying concentrations of

WT hABHD12

transfected HEK293T membrane lysates. We

found a nice linear correlation between the

enzyme concentration, and all three substrates

(Figure 4A). Based on this study we chose 20 µg

membrane lysate for all subsequent studies, as this

gave us a good enzymatic rate in our assays. Next,

we assessed the relation between the enzymatic

rate, and the time of the assay. Like the previous

study we used C18:1 lyso-PS, C18:1 1-MAG, and

4

ABHD12 prefers very long chain lipids

> C20:0 > C18:0 > C16:0 > C14:0 > C12:0 >

C10:0 trend for V_maxand V_max/K_m, and C24:0 <

C22:0 < C20:0 < C18:0 < C16:0 ≈ C14:0 ≈ C12:0

< C10:0 trend for K_m(Figure 5A). Next, when

comparing for 1-MAG lipids for a particular fatty

acid chain length, we do not find any significant

change in kinetic constants with increasing degree

of unsaturation (e.g. for the C20:0, C20:1 and

C20:4 group or C22:0, C22:1, C22:4 and C22:6

group) (Table 1). To complement our 1-MAG

substrate profiling studies, we purchased three

commercially available 2-MAG lipid substrates

with increasing fatty acid chain length and

assayed them against recombinant hABHD12.

Consistent with the enzyme kinetics data from the

1-MAG substrate profiling, we find for 2-MAG

lipids, that hABHD12 prefers C20:4 > C18:1 >

C16:0 from all the kinetic constants for these

lipids (Table 1). We also purchased three

commercially available lyso-PS lipids, and

assayed them against hABHD12. Since, only C16

and C18 fatty acid chain length lyso-PS lipids are

readily available commercially, we find that

hABHD12 prefers, C18:0 and C18:1 lyso-PS

lipids over C16:0 lyso-PS as substrates, and there

isn’t much difference in the kinetic constants

C18:0 and C18:1 lyso-PS (Table 1). Based on all

the kinetic constants from our lipid substrate

profiling, we find that hABHD12 has lyso-PS > 1-

MAG > 2-MAG lipid substrate preference

(Figure 5B), which is consistent with literature

precedence, and within the 1-MAG lipid

substrates, we find that hABHD12 has a very

strong preference for VLC containing 1-MAG

lipids (Figure 5A, Table 1). We would like to

note that we are are not clear what physical form

of the substrate (micellar or free lipid or both)

hABHD12 can hydrolyze, and the kinetic

constants reported in Table 1, are presented as a

function of the total substrate concentration, and

are independent of these physical forms for the

substrates. Finally, we tested the membrane

lysates from S246A hABHD12 against the 1-

MAG lipids at 100 µM, and found no appreciable

activity for the S246A hABHD12 against any 1-

MAG lipid substrate in comparison to wild type

(WT) ABHD12 (Supplementary Figure 1).

Substrate profiling study against endogenous

ABHD12 from mouse brain

To complement the enzyme kinetic

studies that used recombinantly expressed

hABHD12, we wanted to determine the 1-MAG

lipid substrate preference of endogenous

mammalian ABHD12. Towards this, we chose the

mouse brain membrane lysates as a model system,

since previous studies have shown that it

expresses high levels of active ABHD12

(5,10,18,19). In this study, we used wild type and

ABHD12 knockout mice that were littermates (5).

In accordance with previous studies, we

confirmed by gel based ABPP, that our brain

membrane lysate preparations ABHD12 knockout

mice were indeed devoid of any ABHD12

activity, while wild type brain membrane

preparations had robust ABHD12 activity (Figure

6A). Additionally, we confirmed by western blot

analysis, that ABHD12 expression was indeed

absent in ABHD12 knockout mouse brain

membrane lysates (Figure 6A). Having confirmed

the lack of ABHD12 expression and activity in

ABHD12 knockout brain membrane lysates, we

decided to use this as an assay control, to measure

the ABHD12 specific 1-MAG lipase activity of

wild type mouse brain membrane lysates.

Previous studies have also shown that the

mouse brain membrane lysate possess three

enzymes, namely monoacylglycerol lipase

(MAGL), ABHD12 and ABHD6, which can

hydrolyze MAG substrates, hence complicating

the specific contribution that ABHD12 has

towards 1-MAG lipase activity in the mouse brain

membrane lysate (10-12). Fortunately, there are

very selective and potent inhibitors available for

5

ABHD12 prefers very long chain lipids

both MAGL (JZL184: MAGL inhibitor) (20) and

ABHD6 (KT195: ABHD6 inhibitor) (21), which

can be used in tandem in mouse brain membrane

lysates to assess the specific 1-MAG lipase

activity of ABHD12 in this endogenous system.

We therefore treated wild type and ABHD12

knockout brain membrane lysates with both

enriched (>85%) in the microsomal fraction,

which is composed primarily of the endoplasmic

reticulum (ER) (anti-calnexin) (Figure 7A).

Corroborating the cellular fractionation studies

from the mouse brains, we find that in both

Neuro-2a (Figure 7B) and MCF7 cells

(Supplementary Figure 3), ABHD12 is

significantly enriched in the microsomal fraction

(> 90%).

o

JZL184 and KT195 (37 C, 1 h, 1 µM each)

(Supplementary Figure 2), and assayed these

inhibitor treated brain membrane lysates against

the 1-MAG lipid substrate library. Concomitant

with the 1-MAG substrate profile of the

recombinantly expressed human ABHD12, we

find that the endogenous mouse brain ABHD12

also displays the best catalytic activity for VLC

containing 1-MAG lipids in the entire panel of 1-

MAG substrates (Figure 6B).

Second, to complement these organelle

fraction studies and confirm the cellular

localization of mammalian ABHD12, we

performed cellular immunofluorescence assays in

conjunction with high-resolution microscopy in

Neuro-2a cells (Figure 7C). We found from these

studies that ABHD12 had significant overlap with

the ER marker (anti-SEC61A), and hence had

significant ER localization (Figure 7C). The

microscopy based cellular immunofluorescence

studies also showed that ABHD12 had virtually

no localization in the nucleus, Golgi,

mitochondria or with the actin cytoskeleton in

Neuro-2a cells (Supplementary Figure 4).

Similar results were also seen in MCF7 cells

(Supplementary Figure 5,6). Finally, we also

performed similar cellular immunofluorescence

studies in primary peritoneal macrophages derived

from wild type or ABHD12 knockout mice, and

found that ABHD12 was indeed localized to the

ER membrane in primary mouse peritoneal

macrophages (Figure 7D).

Cellular localization of ABHD12

Having established a substrate profile for

both recombinantly expressed and endogenous

ABHD12, we wanted to determine whether there

exists any correlation between the VLC fatty acid

preference for this enzyme and its cellular

localization. To address this, we used two

approaches. First, we fractionated the different

cellular components using established cellular

organelle fractionation techniques and assessed

the ABHD12 localization by western blot

analysis. This cellular fractionation method is

relatively inexpensive, and typically affords

cellular organelle fractions that are ≥ 90%

enriched, and hence provides a tentative idea of

the enrichment of a particular protein in a

particular fraction before venturing into more

sophisticated techniques like microscopy (22). We

performed this fractionation study on wild type

mouse brains, and in two mammalian cell lines

(Neuro-2a and MCF7), since ABHD12 expression

and activity has previously been confirmed in

these mammalian cells (15,21). In mouse brains,

we found that ABHD12 was significantly

Discussion

The integral membrane serine hydrolase

enzyme ABHD12 is expressed highly in the

central

nervous

and

immune

system

(http://biogps.org) (5,15,18,19), and mutations to

this enzyme in humans results in the neurological

disorder, PHARC (1). Recently the murine model

of PHARC, i.e. the ABHD12 knockout mouse

was generated and characterized, and shown to

have massive accumulation of VLC containing

6

ABHD12 prefers very long chain lipids

lyso-PS lipids in the brain, which were correlated

to the age-dependent PHARC like phenotypes

displayed by these mice (5). Here, using a

(K_m). We thus speculate that this might be another

example of an enzyme showing increased

specificity due to improved catalysis, and not

improved substrate binding. Further, we find that

hABHD12 does not have any preference for

unsaturated 1-MAG lipids for a particular chain

length, suggesting that the lipid substrate

preference lies predominantly for the fatty acid

chain length and not the extent of unsaturation of

the 1-MAG substrate (Table 1). We also tested

the endogenous mouse brain ABHD12 against the

1-MAG lipid library, and we find that endogenous

mouse brain ABHD12 also exhibits a similar

substrate profile, as that of recombinant

hABHD12, in that VLC 1-MAG lipids are the

most preferred substrates for ABHD12 (Figure

6). We would also like to add that since we are

performing these substrate assays with membrane

lysates containing the overexpressed enzyme and

not the purified enzyme, we cannot rule out the

possibility that the apparent substrate specificity

of hABHD12 might be due to the increased

solubility of the VLC substrates in the membrane,

that might make these substrates more accessible

to the enzyme. Finally, we show by both cellular

classical

biochemical

approach,

broadly

encompassing synthetic organic chemistry to

make a substrate library, rigorous enzyme kinetics

studies, coupled with cellular fractionation and

immunofluorescence, we set out to provide an

explanation as to why the VLC containing lyso-

PS lipids are massively elevated in the brains of

these mice.

Previous in vitro studies have shown that

ABHD12 accepts 1-MAG lipids as substrates with

catalytic efficiencies quite comparable to lyso-PS

lipids, the putative endogenous substrates of

ABHD12 (Figure 1) (11,12). First, we developed

a facile synthetic scheme for the generation of a

library of 1-MAG lipids from well-known

reactions, and this synthetic route afforded the

generation of 1-MAG lipids of varying chain

length and unsaturation with fairly good yields,

even at a low milligram synthetic scale (Figure

2). We leveraged this synthesis to generate VLC

1-MAG lipids, which have to date, not been

assayed against mammalian ABHD12. We

overexpressed hABHD12 in HEK293T cells, and

show that glycosylation of hABHD12 is critical

for its activity (Figure 3). We also report the

optimization of our LC-MS based lipase assay for

enzyme kinetic measurements, and find a nice

linear correlation between enzyme activity with

enzyme concentration and assay time (Figure 4).

Next, we tested this 1-MAG lipid library

consisting of fatty acids ranging from C10 – C24,

against recombinant hABHD12, and find from the

enzyme kinetics study that hABHD12 has a strong

preference for VLC 1-MAG lipids as substrates

(Figure 5, Table 1). Based on the kinetic

constants (Table 1), it appears that the VLC

substrate specificity for hABHD12 (V_max/K_m) is

likely due to the increased rate of the reaction

(V_max), rather than the decreased apparent affinity

organelle

fractionation

and

cellular

immunofluorescence, that mammalian ABHD12

has a predominant ER membrane localization in

different mammalian cells (including primary

peritoneal macrophages) and the mouse brain

(Figure 7).

In mammals, VLC fatty acids (≥ C22) are

biosynthesized predominantly by an enzyme

named fatty acid elongase (23). This enzyme in

humans, is encoded by the ELOVL genes, has 7

isoforms, and is a membrane bound ER resident

enzyme (23). As a consequence, significant

cellular biosynthesis of VLC fatty acids occurs in

the ER (23,24). In humans, VLC fatty acids serve

several functions, like maintenance of myelin

sheath, normal neuronal function, liver

homeostasis, retinal functions, skin barrier

7

ABHD12 prefers very long chain lipids

formation, anti-inflammatory properties, and

spermatogenesis amongst others (24-26). Not

surprisingly, mutations in genes that are involved

in VLC fatty acid biosynthesis or metabolism,

cause several inherited disorders in humans.

Projecting forward, it would be beneficial

to develop a synthetic route towards making lyso-

PS lipids of varying chain lengths, so that our

current biochemical study can be validated against

the putative endogenous substrate of ABHD12.

Additionally, the new synthetic route might afford

the generation of photoreactive bioorthogonal

lyso-PS probes, which in conjunction with

recently established mass spectrometry based

chemoproteomics (28-30) would greatly facilitate

the discovery of protein ligands, and/or receptors

of the possibly pathological VLC lyso-PS lipids,

that are as of yet unknown. While it is now known

that much of the VLC PS biosynthesis occurs in

the ER (24,27), the biosynthetic origins of VLC

lyso-PS lipids remain elusive. There are currently

two candidate enzymes namely ABHD16A (15)

and PS-PLA1 (31,32) that are responsible for the

biosynthesis of lyso-PS lipids from PS precursors.

Understanding the substrate scope especially the

VLC PS preference and mapping the cellular

localization of these enzymes, would greatly

facilitate the understanding of the biosynthesis of

VLC lyso-PS lipids, and its spatiotemporal

localization in mammalian cells.

Before

genetic

mapping,

PHARC

was

misdiagnosed as Refsum’s disease (or syndrome),

another autosomal recessive genetic disease, as

human subjects with the latter also clinically

present with peripheral neuropathy, retinitis

pigmentosa, hearing loss, early onset of cataract,

and cerebellar ataxia (2). Of note, Refsum’s

disease is caused by mutations to the PHYH gene,

which is responsible for the metabolism of a

dietary branched chain fatty acid, phytanic acid,

through its α-oxidation in the peroxisomes. The

defective metabolism and resulting accumulation

of phytanic acid in the blood and nervous system

tissues due to the faulty PHYH gene function

causes

the

aforementioned

symptoms.

Interestingly, the symptoms of both Refsum’s

disease, and PHARC, are very identical, and

studies from the murine model of PHARC,

suggests that the PHARC like phenotypes might

also be attributed to defective lipid metabolism, in

the central nervous system (lyso-PS in case of

PHARC). Here, we show that ABHD12 is

localized at the cellular compartment (ER

membrane) where the biosynthesis of both VLC

fatty acids, and phosphatidylserine lipids is the

highest (24,27). Our biochemical studies which

includes enzyme kinetics and cellular localization

data, along with the studies from others (5),

suggest speculatively that given ABHD12’s

preference for VLC lipids, it functions as a major

lipase controlling the concentrations and flux of

the VLC lipids (lyso-PS), which are constantly

biosynthesized in the ER membrane, and

perturbation in the activity of ABHD12 causes

unchecked accumulation of these VLC lipids

particularly lyso-PS lipids, resulting in the

pathophysiology observed in PHARC subjects.

Experimental Procedure

Materials

All chemicals, buffers, solvents and

reagents were purchased from Sigma-Aldrich

(now Merck), all lipids and lipid standards were

purchased from Avanti Polar Lipids Inc., and all

primary and secondary antibodies were purchased

from Abcam, unless otherwise mentioned.

Synthesis of 1-MAG lipids

For reaction 1: To a solution of the

desired free fatty acid (1.0 equivalent) and 1,2-

isopropylideneglycerol (1.0 equivalent) in

anhydrous dichloromethane (CH₂Cl₂, 1.5 mL)

maintained

at

0

°C,

N,N-dimethyl-4-

aminopyridine (DMAP, 0.25 equivalent) and 1-(3-

8

ABHD12 prefers very long chain lipids

dimethylaminopropyl)-3-ethylcarbodiimide

hydrochloride (EDC·HCl, 1.0 equivalent) was

sequentially added. The reaction mixture was

warmed to room temperature and stirred for 16 h.

Upon disappearance of starting compound

observed on TLC, the reaction was quenched with

saturated sodium bicarbonate (NaHCO₃) and

extracted three times with CH₂Cl₂. The combined

organic layer was dried over sodium sulfate

(Na₂SO₄), filtered and the filtrate was

concentrated. The crude residue was purified by

column chromatography using 5% ethyl acetate/

hexane as an eluent to afford the corresponding

fatty acid ester.

abbreviations are used: br (broad signal), m

(multiplet), s (singlet), d (doublet), t (triplet) and

dd (doublet of doublets). High-resolution mass

spectra were obtained from HRMS-ESI-Q-Time

of Flight LC-MS/MS (Sciex) available at IISER

Pune Mass Spectrometry Core.

Expression of hABHD12 in HEK293T cells and

preparation of membrane lysates

The full-length WT hABHD12 cDNA

was purchased from GE Life Sciences

Dharmacon, cloned in pCMV-Sport6 vector

between the NotI and SalI restriction sites. The

catalytically inactive S246A hABHD12 mutant

(5) was generated using DpnI based site directed

mutagenesis using Phusion polymerase (New

England Biolabs) as per manufacturers

instructions. HEK293T cells were purchased from

ATCC, and cultured in complete DMEM media

that contained DMEM (HiMedia) supplemented

with 10% (v/v) fetal bovine serum (FBS)

(Thermofisher Scientific), and 1x-penicillin

For reaction 2: Amberlyst-15 (H⁺form,

0.5 equivalent) was added to a solution of fatty

acid ester (1.0 equivalent) in methanol (MeOH).

The resulting reaction mixture was stirred for 16 h

at room temperature. After completion of reaction

(TLC analysis), Amberlyst-15 was filtered off and

the filtrate was evaporated under reduced

pressure. The crude residue was purified by

column chromatography using 40% ethyl acetate/

hexane as an eluent to afford the corresponding 1-

MAG lipid. The detailed synthesis and compound

characterization data of each individual 1-MAG

lipid can be found in the Supplementary

Material.

o

streptomycin (MP Biomedicals) at 37 C with 5%

(v/v) CO₂. The cells were periodically stained

with DAPI, to ensure they were devoid of any

mycoplasma contamination. The hABHD12 was

expressed recombinantly in HEK293T cells using

a previously described transfection protocol (15).

Briefly, HEK293T cells were grown to 35%

confluence in complete DMEM media (15 cm

General material and methods for the

synthesis: Merck silica gel TLC plates (0.25 mm,

60 F254) were used to monitor all chemical

o

dish) at 37 C with 5% (v/v) CO₂, and the cells

reactions.

Column

chromatography

was

were transiently transfected with the above-

mentioned vectors (WT- and S246A-ABHD12)

using polyethyleneimine “MAX” 40,000 MW

(PEI, Polysciences Inc.) (1:3 = plasmid:PEI, 15 µg

plasmid for 15 cm dish). “Mock” control cells

were transfected with an empty vector using the

same protocol (15). The cells were harvested by

scraping 48 h after transfection, washed with

sterile Dulbecco’s phosphate buffer saline without

calcium and magnesium (pH 7.2) (DPBS)

(Thermofisher Scientific) (3x), re-suspended 1

performed using silica gel Rankem (60–120 mesh)

or silica gel Spectrochem (100–200 mesh). Unless

otherwise specified, H and ¹³C spectra were

1

recorded on a JEOL 400 MHz (or 100 MHz for

¹³C) or a Bruker 400 MHz (or 100 MHz for ¹³C)

spectrometer using either residual solvent signals

(CDCl₃δ_H= 7.26 ppm, δ_C= 77.2 ppm) or as an

internal tetramethylsilane (δ_H= 0.00, δ_C= 0.0).

Chemical shifts (δ) are reported in ppm and

coupling constants (J) in Hz. The following

9

ABHD12 prefers very long chain lipids

mL DPBS, and lysed by sonication. The cellular

debris was pelleted by centrifugation at 200g for 5

µL of cold sterile DPBS was added, mixed by

pipetting, centrifuged at 1000g for 5 min at 4 ^oC to

separate the tissue debris. The resulting

supernatant (~700 µL) was separated and

o

min at 4 C, and the resulting lysate (~ 900 µL)

was separated, and centrifuged at 100,000g for 45

o

min at 4 C. The supernatant was discarded, and

centrifuged at 100,000g for 45 min at 4 C. The

the pellet (membrane proteome) was washed with

cold sterile DPBS (3x), and re-suspended in 500

µL of cold sterile DPBS by pipetting. The protein

concentration was estimated using BCA Protein

assay kit (Pierce), and the expression and activity

(lack of it in S246A hABHD12) of hABHD12

was confirmed by western blot analysis, and gel

based ABPP respectively (16). For the gel based

ABPP assays, the membrane proteomes (1

mg/mL, 100 µL) were treated with 2 µM FP-

resulting supernatant was discarded and the pellet

(membrane proteome) was washed with cold

sterile DPBS (3x), and re-suspended in 1 mL of

cold sterile DPBS by pipetting. The protein

concentration of the brain membrane proteome

was estimated using BCA Protein assay kit

(Pierce).

PNGaseF treatments

For the gel based ABPP assays, the

o

rhodamine for 45 min at 37 C with constant

PNGaseF treatments were done under denaturing

conditions as per manufacturers instructions

(Sigma-Aldrich). Briefly, 50 µg of HEK293T

transfected ABHD12 membrane lysates were

treated with FP-rhodamine (4 µM, 1 h, 37 °C, 650

rpm shaking). To this 0.5X of glycoprotein

denaturing buffer containing 5% (w/v) sodium

dodecylsulfate (SDS) and 400 mM dithiothreitol

(DTT) was added and denatured by boiling at 95

^oC for 10 min. The samples were than cooled

instantly on ice for 5 min and 0.1% (w/v) NP-40

was added. To this denatured labeled proteome

PNGaseF at 0.01 U/µg proteome was added and

shaking. Addition of 4X-SDS loading buffer

o

followed by boiling the samples at 95 C for 10

min, quenched these reactions. Fluorescently

labeled proteomes were resolved on a 10% SDS-

PAGE gel and samples were visualized for

enzyme activity by in-gel fluorescence using a

Syngene G-Box Chemi-XRQ gel documentation

system.

Preparation of brain membrane lysates

All mouse studies were performed

following protocols that received approval from

the IISER Pune Institutional Animal Ethics

Committee. The mice were genotyped using an

established protocol (5). The mouse brain

membrane proteomes were prepared using a

previously described procedure (15). Briefly, the

mice were first anaesthetized with isoflurane, and

euthanized by cervical dislocation, following

which the brains of the mice were harvested. A

fresh half brain was suspended in 500 µL of cold

sterile DPBS, and homogenized using a tissue

homogenizer (Bullet Blender 24, Next Advance)

with 1 scoop of glass beads (0.5 mm diameter,

Next Advance) at a speed setting of 8 for 3 min at

o

incubated at 37 C for 1 h with constant shaking

(750 rpm). Reaction was quenched with 4X

loading buffer and proteome was resolved on 12%

SDS-PAGE gel. In-gel fluorescence of labeled

protein was visualized using Syngene G-Box

Chemi-XRQ gel documentation system and

western blot was performed for the same gel. For

LC-MS based lipase assays, the PNGaseF

treatments were done under non-denaturing

conditions. Briefly, the membrane proteome was

treated with PNGaseF at 0.1 U/µg proteome and

o

0.1% (w/v) NP-40 and incubated at 37 C for 1 h

with constant shaking (750 rpm). 20 µg of this

o

4 C. To the brain homogenate, an additional 500

membrane proteome was used for lipase assays as

10

ABHD12 prefers very long chain lipids

described below. All substrates were assayed at

100 µM in this assay.

Western blot analysis

Membrane lysates were resolved on a

10% SDS-PAGE gel, and transferred onto a

PVDF membrane (GE Healthcare) (60 V for 12 h

at 4 ^oC). Post-transfer, the membrane was blocked

with 5% (w/v) milk in Tris buffered saline (pH

7.4) containing 0.1% (w/v) Tween 20 (TBST),

and subsequently probed with a primary antibody

(dilution 1:1000). Thereafter the membrane was

washed with TBST (3x), and incubated with the

appropriate secondary antibody for 90 min at 25

^oC. The membrane was subsequently washed with

TBST (3x) and the signal was visualized using the

Thermo West Pico western blotting substrate

(Thermofisher Scientific) using a Syngene G-Box

Chemi-XRQ gel documentation system. The

primary antibodies used in the study were: anti-

ABHD12 (Rabbit, Abcam, 182011), anti-ATP5A

(Mouse, Abcam, ab14748), anti-GM130 (Mouse,

Abcam, ab169276), anti-Calnexin (Rabbit,

Abcam, ab10286), anti-PMP70 (Mouse, Sigma-

Aldrich, SAB4200181), anti-Lamin (Rabbit,

Cloud Clone Corp, CAF548Hu01), anti-Tubulin

(Rabbit, Cloud Clone Corp, CAB870Hu01), anti-

Lipid substrate assays

All lipase assays were performed with 20

µg of membrane proteome in a final volume of

100 µL. The enzyme kinetics assays were

performed with varying substrate concentrations

(0 – 400 µM), while all single concentration

assays were performed at 100 µM of the

respective lipid substrate. The lipase assays were

performed using a previously established protocol

in glass vials (15,33,34). Briefly, 20 µg of the

membrane proteome was incubated with lipid

substrate of desired concentration in DPBS (100

o

µL total volume) at 37 C with constant shaking.

After 30 min, the reaction was quenched by

addition of 250 µL of 2:1 (v/v) chloroform

(CHCl₃): methanol (MeOH) containing 1.25 nmol

of the internal standard (heptadecenoic acid,

C17:1 FFA). The two phases were separated by

centrifugation at 3000g for 5 min, and the organic

phase (bottom) was removed. The organic extracts

were dried under a stream of N₂, and re-

solubilized in 100 µL of 2:1 (v/v) CHCl₃:MeOH.

The LC-MS analysis was done using a previously

established protocol (15,33,34). Measuring the

area under the curve and normalizing to the

respective internal standard was used to quantify

the free fatty acid release in this lipase assay. The

non-enzymatic rate of substrate hydrolysis was

obtained by using heat-denatured proteomes (15

GAPDH

(Mouse,

Cloud

Clone

Corp,

CAB932Hu01). The secondary antibodies used in

this study were: HRP-conjugated anti-Rabbit IgG

(Goat, Thermofisher Scientific, 31460), or HRP-

linked anti-Mouse IgG (Goat, Cloud Clone Corp,

SAA544Mu19).

Cellular fractionation studies

o

min at 95 C, followed by cooling at 4 C for 10

min, 3 times) as a control, and this value was

subtracted from the substrate hydrolysis rates of

native proteomes to yield the corrected rates. For

the enzyme kinetics assays, the mock membrane

hydrolysis rate was subtracted from the WT

hABHD12 membrane hydrolysis rate for a

particular lipid substrate concentration, to yield a

corrected enzymatic rate for the WT hABHD12

membrane lysates.

The cellular organelle fractionation from

mouse brains and mammalian cell lines were

performed using a previously described protocol

(22). Briefly, mammalian cells from 6 x 15 cm

tissue culture plates (80% confluent) or half-mice

brain was dounce homogenized in 7 mL 250-STM

homogenization buffer containing 250 mM

sucrose, 50 mM Tris-HCl (pH 7.4), 5 mM MgCl₂,

1 mM DTT, 25 µg/mL spermine and 25 µg/mL

spermidine, 1 mM PMSF until floating nuclei

11

ABHD12 prefers very long chain lipids

o

were seen under a phase contrast microscope. The

unlysed cells were pelleted by centrifuging at

200g for 3 min at 4 ^oC, and the resulting

supernatant was centrifuged at 1000g for 15 min

at 4 ^oC to pellet the nuclear fraction. The

supernatant from this step, was further centrifuged

seeded with 2.5 x 10⁶cells, for 24 h at 37 C and

5% (v/v) CO₂. Cells were washed with sterile

DPBS and fixed with 4% (w/v) paraformaldehyde

o

in DPBS for 20 min at 25 C. Subsequently, the

fixed cells were washed with DPBS (3x), and

permealized with 0.5% (v/v) Triton X-100 in

DPBS containing 5% (w/v) bovine serum albumin

(BSA) for 15 min at 25 ^oC. Thereafter the

permealized cells were blocked with 5% (w/v)

BSA in DPBS (blocking buffer) for 30 min at 25

^oC and probed with primary antibody (1:100

o

at 10,000g for 15 min at 4 C to pellet the crude

mitochondrial fraction, and the left over

supernatant was centrifuged at 100,000g for 60

min at 4 °C. The resulting pellet was collected as

microsomal fraction and supernatant was

collected as the cytosolic fraction. The crude

mitochondrial fraction was additionally cleaned

by re-suspending this fraction in 10 volumes of

2M-STMDPS buffer containing 2 M sucrose, 50

mM Tris-HCl (pH 7.4), 5 mM MgCl₂, 1 mM

DTT, 25 µg/mL spermine and 25 µg/mL

spermidine, 1 mM PMSF, and centrifuging it at

o

dilution) for 90 min at 25 C, following which the

secondary antibody treatments (1:1000 dilution)

o

were done for 60 min at 25 C. Lastly the nuclei

were stained with 4’,6-diamidino-2-phenylindole

(DAPI, nuclear marker; Sigma-Aldrich) for 5 min

o

at 25 C. For confocal microscopy: The coverslip

was then mounted on a glass slide with a drop of

FluoromountTM aqueous mounting medium

(Sigma-Aldrich). The mounting medium was

o

6,000g for 15 min at 4 C to pellet a pure

mitochondrial fraction. Each pelleted fraction was

re-suspended in cold sterile DPBS and protein

was quantified using Bradford reagent (Sigma-

Aldrich), following which the proteome was

denatured by boiling with 4X loading buffer at 95

°C for 15 min. The denatured proteomes were

resolved on a 10% SDS-PAGE gel, and assessed

by above described western blot analysis for

enrichment of appropriate proteins in the

respective cellular fraction.

o

allowed to dry in the dark at 25 C and the slides

were imaged on a Zeiss LSM 710 confocal laser-

scanning microscope fitted with a 63X, 1.4 oil

immersion objective. For Stimulated Emission

Depletion (STED) microscopy: the coverslip was

then mounted on a glass slide with a drop of

Mowiol-DABCO mounting media. The mounting

o

medium was allowed to dry in the dark at 25 C

and the slides were imaged on a Leica TCS STED

Super resolution microscope with HC PL APO

CS2, 100X, 1.4 oil immersion objective. The raw

images were deconvoluted using Huygens

Professional Software (Scientific Volume Imaging

BV) and all microscopy images were analyzed

using ImageJ 1.50i software (NIH, for Windows

OS). The primary antibodies used in these studies

were: anti-ABHD12 (Rabbit, Abcam, ab87048 or

Mouse, Abcam, ab68949), anti-KDEL (Mouse,

Abcam, ab12223, ER marker), anti-ATP5A

(Mouse, Abcam, ab14748, Mitochondrial marker),

anti-GM130 (Mouse, BD Transduction Labs,

610822, Golgi marker). Alexa FluorTM 488

Cellular immunofluorescence studies

The cellular immunofluorescence studies

were performed using manufacturer recommended

protocols (Thermofisher Scientific). The Neuro-2a

and MCF7 cell lines were purchased from ATCC,

and cultured in DMEM (HiMedia) with 10% (v/v)

FBS (Thermo Fischer Scientific) and 1%

penicillin-streptomycin (MP Biomedicals). The

cells were counted using a trypan blue method on

a BTC20 automated cell counter (Bio-Rad) as per

manufacturers instructions, and a 35 mm tissue

culture dish containing a 18 mm coverslip, was

12

ABHD12 prefers very long chain lipids

Phalloidin (Thermofisher Scientific, A12379) or

Alexa FluorTM 568 Phalloidin (Thermofisher

Scientific, A12380) was used to stain the actin

cytoskeleton. Goat anti-rabbit IgG (H&L)

conjugated with DyLight® 488 (Thermofisher

Scientific, 35552), or donkey anti-rabbit IgG

(H&L) conjugated with Alexa Fluor® 568

(Thermofisher Scientific, A-10042), or goat anti-

mouse IgG (H&L) conjugated with Alexa Fluor®

Plus 488 (Thermofisher Scientific, A-11017), or

goat anti-mouse IgG (H&L) conjugated with

Alexa Fluor® 568 (Thermofisher Scientific, A-

11004) were used as secondary antibodies.

Data fitting and plotting

The enzyme kinetics data was fit to a

classical Michaelis-Menten equation built in the

Prism 7 software for Mac OS X (GraphPad). All

data was plotted using the same software. All data

are shown as mean ± standard deviation for three

(or more) biological replicates.

13

ABHD12 prefers very long chain lipids

Acknowledgement: Prof. Benjamin F. Cravatt, The Scripps Research Institute, is thanked for

providing ABHD12 knockout mice. Prof. Harinath Chakrapani, IISER Pune is thanked for

providing critical inputs to the manuscript and supervision for all the synthesis reactions

described in this study. We thank Girish Ratnaparkhi, Nagaraj Balasubramanian, Richa Rikhy

and Kundan Sengupta for providing antibodies used in this study. Dr. Ajeet Singh is thanked for

access to the LC-MS facility at CAMS, NCL Venture Center. The National Facility for Gene

Function in Health and disease, IISER Pune, is thanked for maintaining and providing mice for

this study. The IISER Pune – Leica microscopy facility is thanked for providing access to the

microscope needed for the immunofluorescence studies.

Conflict of interest: The authors declare that they have no conflicts of interest with the contents

of this article.

Author contributions: A. J., M. S. and S. S. contributed equally to this work. A. J., S. S., and A.

R. performed the biochemical experiments and analyzed the data. M. S. and A. M. synthesized

and characterized the compounds in this study. S. S. K. conceived the project, and wrote the

manuscript with inputs from all authors.

14

ABHD12 prefers very long chain lipids

References

1.

Fiskerstrand, T., H'Mida-Ben Brahim, D., Johansson, S., M'Zahem, A., Haukanes, B. I.,

Drouot, N., Zimmermann, J., Cole, A. J., Vedeler, C., Bredrup, C., Assoum, M., Tazir,

M., Klockgether, T., Hamri, A., Steen, V. M., Boman, H., Bindoff, L. A., Koenig, M.,

and Knappskog, P. M. (2010) Mutations in ABHD12 cause the neurodegenerative disease

PHARC: An inborn error of endocannabinoid metabolism. Am J Hum Genet. 87, 410-417

Fiskerstrand, T., Knappskog, P., Majewski, J., Wanders, R. J., Boman, H., and Bindoff,

L. A. (2009) A novel Refsum-like disorder that maps to chromosome 20. Neurology 72,

20-27

Chen, D. H., Naydenov, A., Blankman, J. L., Mefford, H. C., Davis, M., Sul, Y., Barloon,

A. S., Bonkowski, E., Wolff, J., Matsushita, M., Smith, C., Cravatt, B. F., Mackie, K.,

Raskind, W. H., Stella, N., and Bird, T. D. (2013) Two novel mutations in ABHD12:

expansion of the mutation spectrum in PHARC and assessment of their functional effects.

Hum Mutat. 34, 1672-1678

2.

3.

4.

5.

Eisenberger, T., Slim, R., Mansour, A., Nauck, M., Nurnberg, G., Nurnberg, P., Decker,

C., Dafinger, C., Ebermann, I., Bergmann, C., and Bolz, H. J. (2012) Targeted next-

generation sequencing identifies a homozygous nonsense mutation in ABHD12, the gene

underlying PHARC, in a family clinically diagnosed with Usher syndrome type 3.

Orphanet journal of rare diseases 7, 59

Blankman, J. L., Long, J. Z., Trauger, S. A., Siuzdak, G., and Cravatt, B. F. (2013)

ABHD12 controls brain lysophosphatidylserine pathways that are deregulated in a

murine model of the neurodegenrative disease PHARC. Proc. Natl. Acad. Sci. U.S.A.

110, 1500-1505

6.

7.

Long, J. Z., and Cravatt, B. F. (2011) The metabolic serine hydrolases and their functions

in mammalian physiology and disease. Chem Rev 111, 6022-6063

Yoshimura, H., Hashimoto, T., Murata, T., Fukushima, K., Sugaya, A., Nishio, S. Y., and

Usami, S. (2015) Novel ABHD12 Mutations in PHARC Patients: The Differential

Diagnosis of Deaf-Blindness. Ann Otol Rhinol Laryngol. 124 Suppl 1, 77S-83S

Nishiguchi, K. M., Avila-Fernandez, A., van Huet, R. A., Corton, M., Perez-Carro, R.,

Martin-Garrido, E., Lopez-Molina, M. I., Blanco-Kelly, F., Hoefsloot, L. H., van Zelst-

Stams, W. A., Garcia-Ruiz, P. J., Del Val, J., Di Gioia, S. A., Klevering, B. J., van de

Warrenburg, B. P., Vazquez, C., Cremers, F. P., Garcia-Sandoval, B., Hoyng, C. B.,

Collin, R. W., Rivolta, C., and Ayuso, C. (2014) Exome sequencing extends the

phenotypic spectrum for ABHD12 mutations: from syndromic to nonsyndromic retinal

degeneration. Ophthalmology 121, 1620-1627

8.

9.

Criscuolo, C., Guacci, A., Carbone, R., Lieto, M., Salsano, E., Nanetti, L., Michele, G.

D., and Filla, A. (2013) Polyneuropathy, hearing loss, ataxia, retinitis pigmentosa and

cataracts (PHARC) screening in an Italian population. Eur J Neurol 20, e60

Blankman, J. L., Simon, G. M., and Cravatt, B. F. (2007) A comprehensive profile of

brain enzymes that hydrolyze the endocannabinoid 2-arachidonoylglycerol. Chem Biol

14, 1347-1356

Savinainen, J. R., Saario, S. M., and Laitinen, J. T. (2012) The serine hydrolases MAGL,

ABHD6 and ABHD12 as guardians of 2-arachidonoylglycerol signalling through

cannabinoid receptors. Acta physiologica 204, 267-276

Navia-Paldanius, D., Savinainen, J. R., and Laitinen, J. T. (2012) Biochemical and

pharmacological characterization of human alpha/beta-hydrolase domain containing 6

(ABHD6) and 12 (ABHD12). J Lipid Res 53, 2413-2424

Christie, W. W. (1993) Preparation of ester derivatives of fatty acids for chromatographic

analysis. in Advances in Lipid Methodology (Christie, W. W. ed.), Oily Press, Dundee. pp

69-111

Iwashita, M., Makide, K., Nonomura, T., Misumi, Y., Otani, Y., Ishida, M., Taguchi, R.,

Tsujimoto, M., Aoki, J., Arai, H., and Ohwada, T. (2009) Synthesis and evaluation of

lysophosphatidylserine analogues as inducers of mast cell degranulation. Potent activities

of lysophosphatidylthreonine and its 2-deoxy derivative. J Med Chem 52, 5837-5863

Kamat, S. S., Camara, K., Parsons, W. H., Chen, D. H., Dix, M. M., Bird, T. D., Howell,

A. R., and Cravatt, B. F. (2015) Immunomodulatory lysophosphatidylserines are

regulated by ABHD16A and ABHD12 interplay. Nat Chem Biol 11, 164-171

10.

11.

12.

13.

14.

15.

15

ABHD12 prefers very long chain lipids

Liu, Y., Patricelli, M. P., and Cravatt, B. F. (1999) Activity-based protein profiling: the

serine hydrolases. Proc Natl Acad Sci U S A 96, 14694-14699

Cravatt, B. F., Wright, A. T., and Kozarich, J. W. (2008) Activity-based protein profiling:

from enzyme chemistry to proteomic chemistry. Annu Rev Biochem 77, 383-414

Wu, C., Jin, X., Tsueng, G., Afrasiabi, C., and Su, A. I. (2016) BioGPS: building your

own mash-up of gene annotations and expression profiles. Nucleic Acids Res 44, D313-

316

16.

17.

18.

19.

20.

21.

Wu, C., Orozco, C., Boyer, J., Leglise, M., Goodale, J., Batalov, S., Hodge, C. L., Haase,

J., Janes, J., Huss, J. W., 3rd, and Su, A. I. (2009) BioGPS: an extensible and

customizable portal for querying and organizing gene annotation resources. Genome Biol

10, R130

Long, J. Z., Li, W., Booker, L., Burston, J. J., Kinsey, S. G., Schlosburg, J. E., Pavon, F.

J., Serrano, A. M., Selley, D. E., Parsons, L. H., Lichtman, A. H., and Cravatt, B. F.

(2009) Selective blockade of 2-arachidonoylglycerol hydrolysis produces cannabinoid

behavioral effects. Nat Chem Biol 5, 37-44

Hsu, K. L., Tsuboi, K., Adibekian, A., Pugh, H., Masuda, K., and Cravatt, B. F. (2012)

DAGLbeta inhibition perturbs a lipid network involved in macrophage inflammatory

responses. Nat Chem Biol. 8, 999-1007

22.

23.

24.

25.

26.

Cox, B., and Emili, A. (2006) Tissue subcellular fractionation and protein extraction for

use in mass-spectrometry-based proteomics. Nature protocols 1, 1872-1878

Jakobsson, A., Westerberg, R., and Jacobsson, A. (2006) Fatty acid elongases in

mammals: their regulation and roles in metabolism. Prog Lipid Res 45, 237-249

Kihara, A. (2012) Very long-chain fatty acids: elongation, physiology and related

disorders. J Biochem 152, 387-395

Sassa, T., and Kihara, A. (2014) Metabolism of very long-chain Fatty acids: genes and

pathophysiology. Biomol Ther (Seoul) 22, 83-92

Lamari, F., Mochel, F., Sedel, F., and Saudubray, J. M. (2013) Disorders of

phospholipids, sphingolipids and fatty acids biosynthesis: toward a new category of

inherited metabolic diseases. Journal of inherited metabolic disease 36, 411-425

Leventis, P. A., and Grinstein, S. (2010) The distribution and function of

phosphatidylserine in cellular membranes. Annual review of biophysics 39, 407-427

Niphakis, M. J., Lum, K. M., Cognetta, A. B., 3rd, Correia, B. E., Ichu, T. A., Olucha, J.,

Brown, S. J., Kundu, S., Piscitelli, F., Rosen, H., and Cravatt, B. F. (2015) A Global Map

of Lipid-Binding Proteins and Their Ligandability in Cells. Cell 161, 1668-1680

Parker, C. G., Galmozzi, A., Wang, Y., Correia, B. E., Sasaki, K., Joslyn, C. M., Kim, A.

S., Cavallaro, C. L., Lawrence, R. M., Johnson, S. R., Narvaiza, I., Saez, E., and Cravatt,

B. F. (2017) Ligand and Target Discovery by Fragment-Based Screening in Human

Cells. Cell 168, 527-541 e529

27.

28.

29.

30.

Hulce, J. J., Cognetta, A. B., Niphakis, M. J., Tully, S. E., and Cravatt, B. F. (2013)

Proteome-wide mapping of cholesterol-interacting proteins in mammalian cells. Nat

Methods 10, 259-264

31.

32.

Aoki, J., Nagai, Y., Hosono, H., Inoue, K., and Arai, H. (2002) Structure and function of

phosphatidylserine-specific phospholipase A1. Biochim Biophys Acta 1582, 26-32

Hosono, H., Aoki, J., Nagai, Y., Bandoh, K., Ishida, M., Taguchi, R., Arai, H., and Inoue,

K. (2001) Phosphatidylserine-specific phospholipase A1 stimulates histamine release

from rat peritoneal mast cells through production of 2-acyl-1-lysophosphatidylserine. J

Biol Chem 276, 29664-29670

33.

34.

Kolar, M. J., Kamat, S. S., Parsons, W. H., Homan, E. A., Maher, T., Peroni, O. D., Syed,

I., Fjeld, K., Molven, A., Kahn, B. B., Cravatt, B. F., and Saghatelian, A. (2016)

Branched Fatty Acid Esters of Hydroxy Fatty Acids Are Preferred Substrates of the

MODY8 Protein Carboxyl Ester Lipase. Biochemistry 55, 4636-4641

Rai, P., Kumar, M., Sharma, G., Barak, P., Das, S., Kamat, S. S., and Mallik, R. (2017)

Kinesin-dependent mechanism for controlling triglyceride secretion from the liver. Proc

Natl Acad Sci U S A 114, 12958-12963

16

ABHD12 prefers very long chain lipids

FOOTNOTES

This research work was supported by a Wellcome Trust-DBT India Alliance Intermediate

Fellowship (IA/I/15/2/502058) to S.S.K., and a DST-FIST Infrastructure Development Grant to

IISER Pune Biology Department.

Abbreviations used in the article:

ABHD12: α/β hydrolase domain containing protein # 12, PHARC: polyneuropathy, hearing loss,

ataxia, retinitis pigmentosa, cataract. , VLC: very long chain, MAG: mono-acyl-glycerol

17

ABHD12 prefers very long chain lipids

Tables

Lipid

V_max

K_m

V_max/K_m

species

(nmol/mg protein/min)

(µM)

(nmol/mg protein/min/M)

1-MAG species

144 ± 14

129 ± 21

119 ± 22

117 ± 21

103 ± 19

106 ± 15

109 ± 15

91 ± 11

C10:0

C12:0

C14:0

C16:0

C18:0

C18:1

C18:3

C20:0

C20:1

C20:4

C22:0

C22:1

C22:4

C22:6

C24:0

C24:1

0.14 ± 0.02

0.16 ± 0.03

0.38 ± 0.04

3.2 ± 0.2

(9.7 ± 0.9) x 10²

(1.2 ± 0.3) x 10³

(3.4 ± 0.4) x 10³

(2.7 ± 0.4) x 10⁴

(5.0 ± 0.5) x 10⁴

(5.1 ± 0.6) x 10⁴

(4.9 ± 0.5) x 10⁴

(1.3 ± 0.2) x 10⁵

(1.4 ± 0.3) x 10⁵

(1.4 ± 0.4) x 10⁵

(2.2 ± 0.3) x 10⁵

(2.0 ± 0.4) x 10⁵

(1.9 ± 0.3) x 10⁵

(1.8 ± 0.3) x 10⁵

(2.7 ± 0.3) x 10⁵

(2.9 ± 0.4) x 10⁵

5.1 ± 0.3

5.6 ± 0.4

5.4 ± 0.3

12.0 ± 0.8

12.3 ± 0.9

12.6 ± 0.9

15.7 ± 1.3

15.6 ± 1.2

15.1 ± 0.8

14.9 ± 1.4

17.7 ± 1.5

17.9 ± 1.8

86 ± 14

91 ± 12

72 ± 12

75 ± 13

78 ± 8

79 ± 9

66 ± 9

61 ± 8

2-MAG species

148 ± 31

129 ± 24

95 ± 17

C16:0

C18:1

C20:4

2.3 ± 0.4

3.6 ± 0.5

6.8 ± 0.5

(1.5 ± 0.3) x 10⁴

(2.8 ± 0.3) x 10⁴

(7.2 ± 0.8) x 10⁴

Lyso-PS species

87 ± 14

C16:0

C18:0

C18:1

7.5 ± 0.7

14.8 ± 1.4

14.3 ± 0.8

(8.6 ± 0.8) x 10⁴

(2.0 ± 0.3) x 10⁵

(1.9 ± 0.3) x 10⁵

73 ± 12

74 ± 11

Table 1: Kinetic constants for the various lipid substrates tested in vitro against recombinant

hABHD12.

18

ABHD12 prefers very long chain lipids

Figure 1: The in vitro substrates and the lipase reactions catalyzed by mammalian ABHD12.

In each case, ABHD12 releases a free fatty acid (FFA) product from the lipid substrate it

hydrolyzes. For lyso-PS lipids, the other product is glycerophosphoserine (GPS), while for 1-

MAG and 2-MAG, the other product is glycerol.

19

ABHD12 prefers very long chain lipids

Figure 2: Facile synthetic route for generating 1-MAG lipids. 1-MAG lipids were generated

using a two-step synthesis. The first step was an EDC coupling reaction to esterify the fatty acid

of interest to a protected glycerol moiety, and the second step, a deprotection reaction using

Amberlyst catalyst to yield 1-MAG lipids. The corresponding yields of each reaction are

tabulated in the figure.

20

ABHD12 prefers very long chain lipids

Figure 3: Glycosylation is needed for optimal hABHD12 activity. (A) The activity of WT

hABHD12, but not S246A hABHD12, was confirmed by gel based ABPP assays on membrane

lysates of HEK293T transfected with WT or S246A hABHD12. The expression over mock of

both WT and S246A hABHD12 was confirmed by western blot analysis. GAPDH was used as a

loading control in these experiments. This experiment was done in biological triplicate with

reproducible results. (B) PNGaseF treatment confirms the glycosylation (migration to a lower

molecular weight band) of overexpressed WT hABHD12 by gel based ABPP, and western blot

analysis. For the gel ABPP assays, the membrane lysates from HEK293T cells transfected with

WT hABHD12 were labeled with FP-rhodamine (2 µM, 1 h, 37 °C), denatured using detergents

21

ABHD12 prefers very long chain lipids

and heating and subsequently treated with PNGaseF (0.01 U/µg proteome) (or DPBS), and in-gel

fluorescence was used as a readout for WT hABHD12. This experiment was done in biological

triplicate with reproducible results. (C) Deglycosylation by PNGaseF results in loss of lipase

activity of hABHD12. Recombinant WT hABHD12 was treated with detergent (0.1% (w/v) NP-

40) or detergent + PNGaseF (0.1% (w/v) NP-40 + 0.1 U/µg proteome), and assayed against C18:1

lyso-PS and C18:1 1-MAG substrates (100 µM each substrate). PNGaseF treated WT hABHD12

has almost no lipase activity against both substrates, comparable to mock and S246A hABHD12

membrane lysate controls. Data represents mean ± standard deviation for three biological

replicates per group.

22

ABHD12 prefers very long chain lipids

Figure 4: Optimization of WT hABHD12 lipase activity assay. (A) The lipase activity of WT

hABHD12 transfected HEK293T membrane lysates vary linearly with lysate concentrations (0 –

50 µg). (B) The lipase activity of WT hABHD12 transfected HEK293T membrane lysate (20 µg)

vary linearly with time of assay (0 – 1 h). All assays were done against three substrates: C18:1

lyso-PS, C18:1 1-MAG and C18:1 2-MAG at 100 µM concentration of each substrate as per the

lipase assay protocol described in the Methods section. Data represents mean ± standard

deviation for 4 biological replicates per group.

23

ABHD12 prefers very long chain lipids

Figure 5: Enzyme kinetic studies with recombinant hABHD12. Enzyme kinetic assays of

hABHD12 with (A) saturated 1-MAG lipids of varying chain length, and (B) C18:1 containing

lyso-PS, 1-MAG and 2-MAG substrates. The data presented in (A) shows that hABHD12 has a

strong preference for VLC 1-MAG lipids, while the data presented in (B) shows that the substrate

preference of hABHD12 is lyso-PS > 1-MAG > 2-MAG. The assays were performed with

membrane lysates from mock or WT hABHD12 transfected HEK293T cells. The mock

membrane lysate background enzymatic rate was subtracted from the WT hABHD12 membrane

lysate rate; to yield a corrected value for hABHD12 enzymatic rate, which is represented in this

enzyme kinetics plot. Each concentration represents rates calculated from three biological

replicates. Data is plotted as mean ± standard deviation for each concentration, and the lines

connecting the points, represents a fit to a classical Michaelis-Menten enzyme kinetics equation.

24

ABHD12 prefers very long chain lipids

Figure 6: Enzymatic assays of 1-MAG lipids with endogenous mouse brain ABHD12. (A)

Gel based ABPP and western blot analysis were used to confirm the loss of ABHD12 activity and

expression in the membrane lysates of brain membrane preparations from wild type (+/+), and

ABHD12 knockout (–/–) mice respectively. As a loading control GAPDH was used. This

experiment was done in biological triplicates with reproducible results. (B) Brain membrane

lysates from wild type (+/+), and ABHD12 knockout (–/–) mice were treated with 1 µM each of

JZL184, and KT195, and were assayed against the 1-MAG lipid library (100 µM of each 1-MAG

substrate). The data represents mean ± standard deviation for three biological replicates per

group.

25

ABHD12 prefers very long chain lipids

Figure 7: Cellular organelle fractionation and immunofluorescence confirm the ER

localization of mammalian ABHD12. Cellular organelle fraction studies showing the

microsomal enrichment of mammalian ABHD12 in (A) the mouse brain, and (B) Neuro-2a cells.

All the cellular fractionation experiments were done in at least biological triplicates with

reproducible results. (C) High resolution cellular immunofluorescence studies in Neuro-2a cells,

showing that ABHD12 has strong co-localization with SEC61A, an ER-marker (yellow in

26

ABHD12 prefers very long chain lipids

merged image). (D) Cellular immunofluorescence studies in primary peritoneal macrophages

from wild type (+/+), and ABHD12 knockout (–/–) mice, showing that ABHD12 has strong co-

localization with KDEL, an ER-marker, in the wild type mice (yellow in merged image in (+/+)

panels), and the lack of ABHD12 expression in the knockout mice (no green fluorescence in (–/–)

panels). All the cellular immunofluorescence experiments were done in at least 4 biological

replicates with reproducible results.

27

Biochemical characterization of the PHARC associated serine hydrolase ABHD12

reveals its preference for very long chain lipids

Alaumy Joshi, Minhaj Shaikh, Shubham Singh, Abinaya Rajendran, Amol Mhetre and

Siddhesh S Kamat

J. Biol. Chem. published online September 20, 2018

Access the most updated version of this article at doi: 10.1074/jbc.RA118.005640

Alerts:

• When this article is cited

• When a correction for this article is posted

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

Article Doi

Biochemical characterization of the PHARC-associated serine hydrolase ABHD12 reveals its preference for very-long-chain lipids

DOI: 10.1074/jbc.RA118.005640

Source and publish data:

Authors:

Article abstract of DOI:10.1074/jbc.RA118.005640

Full text of DOI:10.1074/jbc.RA118.005640

Products guided by the article

R&D Labs maybe for 93894-92-9

Relevant to this article

Hot Product