Deficient Activity of the Fatty Aldehyde Dehydrogenase Component
ofFattyAlcohol:NAD+ Oxidoreductase in Cultured Fibroblasts
William B. Rizzo and Debra A. Craft
Departments ofPediatrics andHuman Genetics, Medical College of
Virginia, Virginia Commonwealth University, Richmond, Virginia 23298
Sjogren-Larsson syndrome (SLS) is an inherited disorder asso-
ciated with impaired fatty alcohol oxidation due to deficient
activity of fatty alcohol:NAD+ oxidoreductase (FAO). FAO is
a complex enzyme which consists oftwo separate proteins that
sequentially catalyze the oxidation offatty alcohol to fatty alde-
hyde and fatty acid. To determine which enzymatic component
ofFAO was deficient in SLS, we assayed fatty aldehyde dehy-
drogenase (FALDH) and fatty alcohol dehydrogenase in cul-
tured fibroblasts from seven unrelated SLS patients. All SLS
cells were selectively deficient in the FALDH component of
FAO, and had normal activity of fatty alcohol dehydrogenase.
The extent ofFALDH deficiency in SLS cells depended on the
aliphatic aldehyde used as substrate, ranng from62% ofmean
normal activity using propionaldehyde as substrate to 8% of
mean normal activity with octadecanal. FALDH activity in ob-
ligate SLS heterozygotes was partially decreased to 49±7% of
mean normal activity using octadecanal as substrate. Differen-
tial centrifugation studies in fibroblasts indicated that this
FALDH enzyme was largely particulate; solubleFALDH activ-
ity was normal in SLS cells. Intact SLS fibroblasts oxidized
octadecanol to fatty acid at < 10% of the normal rate, but oxi-
dized free octadecanal normally, suggesting that the FALDH
affected in SLS is chiefly involved in the oxidation of fatty
alcohol to fatty acid. These results show that the primary enzy-
matic defect in SLS is the FALDH component of the FAO
complex, which leads to deficient oxidation of fatty aldehyde
derived from fatty alcohol. (J. Clin. Invest. 1991. 88:1643-
1648.)Keywords: ichthyosis-mental retardation*ipidmetabo-
lism * genetic disease-neurological disease
Sjogren-Larsson syndrome (SLS)' is an autosomal recessive
disorder characterized by the presence ofcongenital ichthyosis,
mental retardation, and spastic di- and tetraplegia (1). The dis-
Address correspondence to William B. Rizzo, M.D., Medical College
of Virginia, P.O. Box 259, MCV Station, Richmond, VA 23298.
Receivedforpublication 22March 1991 andin revisedform 22May
1. Abbreviations used in this paper: FADH, fatty alcohol dehydroge-
nase; FALDH, fatty aldehyde dehydrogenase; FAO,
hol:NAD+ oxidoreductase; SLS, Sjogren-Larsson syndrome.
order was first described more than 30 years ago in a group of
Swedish patients, but additional cases have been reported
SLS is one of several ichthyotic syndromes associated with
abnormal lipid metabolism. Cultured skin fibroblasts and leu-
kocytes from SLS patients were recently found to have defi-
cient activity offatty alcohol:NAD' oxidoreductase (FAO), an
enzyme that catalyzes the oxidation of fatty alcohol to fatty
acid (5, 6). This biochemical defect results in accumulation of
fatty alcohol in cultured fibroblasts (5) and the plasma (6) of
most SLS patients.
FAO is a complex enzyme consisting of separate proteins
that sequentially metabolize fatty alcohol to fatty aldehyde and
fatty acid, reactions which are catalyzed by fatty alcohol dehy-
(FALDH), respectively (7). Consequently, the biochemical de-
fect in SLS could involve either component ofthe FAO com-
plex. We now report that SLS patients are specifically deficient
in the FALDH component ofFAO.
Chemicals. [1-'4C]Palmitate (58 mCi/mmol), [I-'4C]stearate (56 mCi/
mmol), and other '4C-labeled fatty acids were obtained from ICN Ra-
diochemicals, Irvine, CA. [1-'4C]-labeled fatty alcohols were synthe-
sized from the corresponding radioactive fatty acids by reduction with
Li(Al)H4 as described (8). Nonradioactive fatty alcohols were obtained
from Sigma Chemical Co., St. Louis, MO. [I-'4C]-labeled fatty alde-
hydes and unlabeled octadecanal and hexadecanal were synthesized
from the corresponding radioactive or nonradioactive fatty alcohols by
reaction with l-chlorobenzotriazole asdescribed (9). Octanal and tetra-
decanal were obtained from Aldrich Chemical Co., Inc., Milwaukee,
WI; other aldehydes were purchased from Sigma Chemical Co.
Nonradioactive fatty aldehydes werequantitated by gas-liquidchroma-
tography (6). All fatty aldehydes were stored in ethanol under a nitro-
gen atmosphere at -200C. Solvents were either reagent grade orHPLC
gradefrom J. T. BakerChemical Co., Phillipsburg, NJ. All otherchemi-
cals were from Sigma Chemical Co.
Cells. Human cultured skin fibroblasts were derived from normal
subjects or SLS patients by skin punch biopsy after obtaining informed
consent. Fibroblast cell lines were grown from seven unrelated SLS
homozygotes from the following countries oforigin: United States (AB,
CB); Chile (AZ); New Zealand (AC); Australia (DE); Sweden (LN);
France (YL). Fibroblast cultures were also obtained from five obligate
SLS heterozygotes, who were parents ofthe SLS homozygotes studied.
Cells were grown at 37°C in an atmosphere of5% CC2, 95% air in
Dulbecco's minimal essential medium supplemented with 10% heat-
inactivated fetal bovine serum, penicillin (100 U/ml), and streptomy-
cin (100Ag/ml).All experiments were performed on confluent cells at
passage 2 through 12.
Enzyme assays. Confluent fibroblasts from one 75-cm2 culture
flask were collected by trypsinization and washed three timeswith PBS.
The cell pellet was homogenized with a glass Teflon motor-driven ho-
mogenizer in I ml of25 mM Tris-HCl, pH 8.0, 0.25 M sucrose.
FattyAldehyde Dehydrogenase Deficiency in Sjtigren-Larsson Syndrome
J. Clin. Invest.
© The American Society for Clinical Investigation, Inc.
Volume 88, November 1991, 1643-1648
FAO activity was measured as described (6). To measure the radio-
active hexadecanal accumulated during the FAO assay, fibroblast ho-
mogenates containing 10-25jigprotein were incubated understandard
FAO assay conditions (6) in the presence of 12 MM [14C]-hexadecanol
for 30 min at 370C. Reactions were terminated by the addition of2 ml
each ofhexane, water, and 0.3 MNaOH in ethanol. After vortexing the
mixture for 1 min, the upper hexane layer containing radioactive hexa-
decanal was removed and the lower phase was reextracted with 2 ml
hexane. The hexane extractswere combined, dried under nitrogen, and
radioactive hexadecanal was purified by thin-layer chromatography on
silica gel G plates using a solvent system consisting ofhexane/chloro-
form/methanol (73/25/2). After addition of 1 ml of 2 N HCl to the
remaining lower phase ofthe reaction mixture, radioactive hexadecan-
oic acid was extracted twice with hexane and isolated by chromatogra-
phyon plates using a solvent system consisting ofhexane/diethyl ether/
acetic acid (60/40/0.9). Hexadecanal, hexadecanol, and hexadecanoic
acid standards (25 ug) were included on each chromatography plate.
The plates were stained with rhodamine G and the lipid spots were
visualized under UV light. The spots containing radioactive hexade-
canoic acid and hexadecanal were collectedbyscrapingthe appropriate
area ofsilica gel from their respective thin-layer chromatography plates
and quantitating by scintillation spectroscopy.
FALDH was assayed fluorometrically by measuring the fatty alde-
hyde-dependent production ofNADH. The reaction was monitored in
a fluorometer ( 1 1; Turner Designs, Sunnyvale, CA) equipped with a
heated microcuvette chamber maintained at 370C. The excitation
wavelength was 365 nm and the emission wavelength was monitored at
460 nm. Reaction tubes contained 50 mM glycine-NaOH buffer, pH
9.5, 1.5 mM NAD+, 0.2 mg of fatty acid-free BSA, 10 mM pyrazole,
and 10-40 Mg homogenate protein in a final vol of 0.4 ml. Reaction
tubes were preheated for 2 min at 370C, and the assay was initiated by
the addition of3Mul ofa 21.5-mM hexadecanal (or octadecanal, tetrade-
canal, or dodecanal) solution in ethanol to achieve a final fatty alde-
hyde concentration of 160 MM. Control incubations consisted ofiden-
tical reaction mixtures, but the reaction was initiated by addition of
ethanol lacking fatty aldehyde. The reactions were typically monitored
for 15 min using a chart recorder. The fatty aldehyde-dependent activ-
ity was calculated by subtracting the change in fluorescence measured
in the absence offatty aldehyde from that measured in the presence of
Aldehyde dehydrogenase activity using propionaldehyde, hexanal,
and octanal as substrates could not be assayed under the same reaction
conditions used for hexadecanal, because these shorter substrates
caused an increase in reaction fluorescence even without added cell
homogenate. Instead, activity was measured using 25 mM Tris-HCl,
pH 8.8 in place of the glycine-NaOH buffer, and the final substrate
concentrations were 0.70 mM, 0.70 mM, and 1.3 mM for hexanal,
octanal, and propionaldehyde, respectively.
FADH activity was assayed in the reverse direction in a reaction
mixture consisting of 50 mM glycine-NaOH, pH 9.5, 0.1 mg fatty
cpm dissolved in 3 Ml ethanol) and 10-25 Mg homogenate protein in a
total vol of0.2 ml. Reactions were initiatedby addition ofthe homoge-
nate. Control incubations lacked homogenate. After 15 min at 370C,
reactions were terminated by addition of 2 ml each of hexane, 0.3 M
NaOH in ethanol, and water. Reaction mixtures were agitated with a
vortex mixer for I min and the upper hexane phase containing radioac-
tive hexadecanol was removed. The lower phase was extracted again
with 2 ml hexane. The hexane extracts were combined and dried under
nitrogen. Radioactive hexadecanol was separated from[14C]-hexade-
canal by chromatography on silica gel G plates using a solvent system
consisting of hexane/chloroform/methanol (73/25/2). Standard hexa-
decanal and hexadecanol were spotted on each plate. The fatty alcohol
region was visualized underUV light after staining the plate with rho-
damine G, and the radioactive hexadecanol was collected by scraping
and quantitated by scintillation spectroscopy. FADH activity was cal-
culated by subtracting the radioactivity measured in control reactions
from that measured in reactions containing fibroblast homogenate,
1 mM NADH, 14 MM['4C]-hexadecanal(- 260,000
and specific activitywasexpressedaspicomoles perminutepermilli-
Fibroblastfatty aldehyde andfatty alcohol oxidation. Cultured fi-
broblasts weregrowntoconfluencyin 35-mm diameter culture dishes.
One daybefore study,the cells were fed with fresh DMEMcontaining
10% fetal bovine serum. The culture medium was removed on theday
of the experiment and replacedwith identical mediumcontainingei-
ther 20 MM [1-'4C]octadecanal or 3.5 MM[1-'4C]octadecanol.After
incubation for 20 min (octadecanal)or 40 min(octadecanol)at370C,
the dishes were placedon ice and the medium wasquicklyremoved.
Cell monolayers were washed twice with ice-cold PBS and cells were
collectedby scrapingin methanol.Lipidswere extractedovernightwith
chloroform/methanol (1/1).Insoluble cellular material waspelletedby
centrifugation and the organicsolvent containingextractedlipidswas
dried undernitrogen. Lipidswereresuspendedin 2 ml of0.3 MNaOH
in 95% ethanol andsaponifiedfor 1 hat 80'C. 2 ml ofwaterwasadded,
and nonsaponifiable lipidswere extracted with hexane and discarded.
After addition of 1 ml of 2 NHCl, saponifiable lipidswere extracted
into hexane and dried. Radioactive fattyacids were isolatedbythin-
layer chromatography on silica gel plates usinga solventsystemcon-
sisting of hexane/ether/acetic acid (60/40/1). Nonradioactive palmi-
tate served as a standard. Fattyacidspotswere visualizedby spraying
the platewith rhodamine G and examined under UVlight. Fattyacid
spots were scraped and radioactivitywas determinedbyscintillation
Fibroblast subcellularfractionation by differential centrifugation.
Confluent fibroblasts from three 75-cm2 culture flasks were collected
by trypsinizationand washed twice with PBS. Cells wereresuspended
in 5.8 ml of 25 mMTris-HCI,0.25 Msucrose, pH 8.0,andhomoge-
nized with 15 strokes in a motor-drivenglass-Teflon homogenizer.All
centrifuged for 10 min at 500gin arefrigerated centrifuge (RC2B;
Sorvall Instruments Div., DuPont Co., Newton, CT) usingan SS-34
rotor. The 500g pelletwasresuspendedin 1 ml ofthe samehomogeni-
zation buffer. The 500g supernatantwassequentially centrifugedfor
30 min at 33,000 gand 120,000 gin anultracentrifuge (L5-50;Beck-
man Instruments, Inc., Palo Alto, CA) using an SW50.1 swinging
bucket rotor. In eachcase,thepelletsandaliquotsofthesupernatants
were removed at each stepforenzyme assay.
The oxidation offattyalcohol tofattyacidproceeds througha
fatty aldehyde intermediate,andrequiresthesequentialaction
of the FADH and FALDHcomponentsofthe FAO complex.
To determine which componentofFAO was deficient in SLS
fibroblasts,we reasoned that a defect in the FADHcomponent
would prevent the formation offatty aldehyde intermediate,
whereas a deficiency in FALDH activitywould lead to accu-
mulation of the fatty aldehyde intermediate. We, therefore,
quantitatedthe amount ofradioactive hexadecanal that accu-
mulated when cell homogenateswere incubated with[I-_4C]
hexadecanol under standard FAO assayconditions. As shown
in Table I, production of radioactive hexadecanoic acid was
deficient in cellhomogenatesfrom five unrelated SLSpatients,
but [1-14C]hexadecanal accumulated almost fivefold more in
the mutant cells than in normal homogenates.These results
pointedto a defect in the FALDHcomponentofFAO.
To confirm these indirect results, we assayedboth enzy-
matic componentsofFAO. FALDH wasassayedfluorometri-
cally by monitoringthehexadecanal-dependent productionof
NADH. Using normal fibroblast homogenates,hexadecanal-
dependentNADHproductionincreasedlinearlyover time for
at least 15 min. As shown inFig. 1,the reaction rate wasopti-
mal at pH 9.5 and was linearly dependent ontheamount of
W B. Rizzo and D. A. Craft
Table I. Radioactive Products Formed during Assayfor Fatty
Alcohol:NAD' Oxidoreductase in Cultured Skin Fibroblasts
Using[J-"C]Hexadecanol as Substrate
acid to hexadecanal
syndrome (n = 5)
(n = 5)
*Radioactive products are expressed as picomoles per minute per
homogenate protein added. The enzyme activity was saturable
with respect to hexadecanal and NAD' concentrations, with
apparent Km values of 16 MM and 100,M, respectively. Little
or no activity was detected when NAD' was replaced with
FADH activity in fibroblasts could not be reliably mea-
sured in the forward direction by monitoring fatty aldehyde
production from ['4C]-hexadecanol due to the large amount of
FALDH activity, which tended to oxidize fatty aldehyde to
fatty acid; conditions to selectively inhibit FALDH without
losing FADH activity were not found. However, FADH could
be measured in the reverse direction by monitoring the
NADH-dependent production of ['4C]-hexadecanol when fi-
broblast homogenates were incubated with radioactive hexade-
canal. Under these conditions, the reaction proceeded linearly
overtime for at least 40 min and wasdependent on the amount
ofhomogenate protein added (Fig. 2). FADH activity was satu-
rated with increasing hexadecanal and NADH concentrations.
The apparent Km for hexadecanal was 3 AM, and the enzyme
displayed two apparent Km's forNADH at 36 AM and 100 ,uM.
FADH activity showed a broad pH dependence between pH 8
and pH 10 (not shown). When NADH was replaced with
NADPH, hexadecanal reduction was detected at 50-60% of
that measured with NADH. The activity measured in the pres-
ence ofboth NADH and NADPH was additive, suggesting the
presence of NADPH-dependent aldehyde reductase. There-
fore, theFADH component was routinely assayed withNADH
The FAO complex and its FALDH and FADH compo-
nents were assayed in fibroblast homogenates from normal
controls and seven unrelated SLS patients using 16-carbon
substrates (Table II). In normal cells, FALDH activity was at
least 30-fold greater than either the total FAO complex activity
or FADH activity, suggesting that the initial step in fatty alco-
hol oxidation (to aldehyde) was rate-limiting for theFAO com-
plex. As shown in Table II, FALDH activitywas deficient in all
SLS fibroblasts, whereas FADH activity was normal.
8.0 8.5 9.0 9.5 10.0
Figure 1. Reaction parameters for FALDH activity in normal fibro-
blast homogenates. Data points represent the mean of duplicate de-
terminations. Except as noted, all reactions were performed at pH 9.5
for 15 min, and contained 160MM hexadecanal and 1.5 mM NAD'.
(A) Reaction pH dependence. Each reaction was performed in glycine
buffer and contained 29 ug homogenate protein. (B) Homogenate
protein dependence. (C) Hexadecanal concentration dependence.
Each reaction contained 29 ug homogenate protein. (D) Effects of
varying NAD' concentration. Each reaction contained 22
10 20 30
10 12 14 16
0.51.0 1.5 2.0 2.5
Figure 2. Reaction parameters for FADH activity in normal fibroblast
homogenates. Data points represent the results ofduplicate determi-
nations. Except as noted, all incubations were for 15 min in the pres-
ence of 14,M hexadecanal and 1 mM NADH. (A) Time course of
the reaction. Each reaction contained 15,ghomogenate protein. (B)
Homogenate protein dependence. (C) Hexadecanal concentration
dependence. Each reaction contained 17jghomogenate protein. (D)
Effects ofvarying NADH concentration. Each reaction contained 20
Mg homogenate protein.
Fatty Aldehyde Dehydrogenase Deficiency in Sjogren-Larsson Syndrome
Table II. Enzyme Activities Associated with Fatty Alcohol
Oxidation in Cultured Skin Fibroblasts from Normal Controls
and Seven Unrelated SLS Patients. Cells Were Assayedforthe
FAO Complex, FALDH, andFADH Using 16-Carbon Substrates
(n = 18)
(n = 10)
(n = 9)
*All enzyme activities are expressed as picomoles per minute per
milligram protein (mean±SD).
The extent of enzyme deficiency in SLS fibroblasts was
more profound with longer chain substrates than with shorter
ones (Fig. 3). Mean FALDH activity in SLS ranged from 27%
ofmean normal activity using dodecanal as substrate to 8% of
mean normal activity with octadecanal. A similar trend was
seen with FAO complex activity, and there was a strong corre-
lation (r = 0.96) between FALDH and FAO activities using
18-carbon substrates. SLS fibroblasts were also deficient in
FALDH activity using shorter substrates, such as octanal (25%
of normal activity) and hexanal (30%), but were only mildly
deficient using propionaldehyde as substrate (Table III). Disul-
firam (10 mM), an inhibitor of acetaldehyde dehydrogenase
(10), had little or no inhibitory effect on FALDH activitywhen
octadecanal was used as substrate (data not shown).
FALDH activity was assayed in fibroblast homogenates
from five unrelated obligate SLS heterozygotes using octade-
canal as substrate. As shown in Table IV, mean FALDH activ-
ity was decreased in SLS heterozygotes to 49±7% ofmean nor-
The subcellular distribution ofFALDH was investigated in
normal andSLS fibroblasts by fractionating cellsby differential
centrifugation. In normal fibroblasts (n = 4), FALDH activity
was found in all fractions, including a 500 g pellet, 33,000 g
pellet, 120,000 g pellet, and a 120,000 g supernatant with a
mean distribution of29±13%, 53±11%, 11±2%, and 7±2% of
total activity, respectively. In SLS cells (n = 3), the residual
FALDH activity showed an abnormal distribution with
21±13%, 53±7%, 1±1%, and 25±7% oftotal activity found in
the 500 gpellet, 33,000gpellet, 120,000 gpellet, and a 120,000
g supernatant fractions, respectively. Except for the 120,000 g
supernatant fraction which showed no decrease in FALDH ac-
tivity, the specific activity ofFALDH (measured as picomoles
per minute per milligram protein) was decreased in all subcel-
lular fractions in SLS cells compared to normal controls (Fig.
4). The greatest deficiency ofFALDH was seen in the 120,000g
pellet fraction(2% ofmean normal activity). These results indi-
cate that the FALDH which is deficient in SLS is a particulate
or membrane-bound enzyme.
Substrate Chain Length
Figure 3. Residual activity of FAO and FALDH in crude homoge-
nates ofSLS fibroblasts using different chain length saturated fatty
alcohols and fatty aldehydes as substrates. Seven SLS cell lines were
assayed and data are expressed as the percentage±SEM ofthe mean
specific activities measured in normal fibroblasts. FAO activities
measured in normal fibroblasts were 142±35 (n=8), 129±30 (n=8),
101±29 (n=18), and 75± 13 (n=14) pmol/min/mg protein using
dodecanol, tetradecanol, hexadecanol, and octadecanol, respectively,
as substrate. FALDH activities measured in normal fibroblasts were
10,622±2,045 (n=9), 9,510±2,326 (n=9), 9,139±2,578 (n=9), and
8,880± 1,084 (n=12) pmol/min/mg protein using dodecanal, tetra-
decanal, hexadecanal, and octadecanal, respectively, as substrate.
Open bars, FAO; cross-hatched bars, FALDH.
The presence of such a profound deficiency of FALDH
activity in SLS fibroblasts raised the possibility that this
FALDH functioned not only as a component ofFAO but also
to oxidize free fatty aldehydes. To determine whether free fatty
aldehyde oxidation was impaired under more physiological
conditions, intact fibroblasts were incubated with ['4C]-octade-
canal and the production ofradioactive octadecanoic acid was
measured. As shown in Fig. 5, oxidation of free octadecanal
was not impaired in SLS fibroblasts. In contrast, oxidation of
Table III. Aldehyde Dehydrogenase Activity in Normal and SLS
Fibroblasts Using Various Aldehyde Substrates
Aldehyde dehydrogenase activity*
*Enzyme activity is expressed as picomoles per minute per milligram
ofprotein (mean±SD) for five normal and five SLS cell lines.
W. B. Rizzo and D. A. Craft
Table IV. Fatty Aldehyde Dehydrogenase Activities in Cultured
Fibroblastsfrom SLS Homozygotes, Obligate SLSHeterozygotes,
and Normal Controls using Octadecanal as Substrate
Obligate SLS Heterozygotes
Normal controls (n = 12)
*FALDH activity is expressed as picomoles per minute per milligram
protein (mean±SD). Each cell line was tested on two to four separate
radioactive octadecanol to fatty acid was decreased in intact
SLS cells to < 10% ofmean normal activity. Unlike the assays
with fibroblast homogenates, however, there was no concomi-
tant cellular accumulation of free radioactive octadecanal in
the intact SLS cells (data not shown).
Figure 4. Differential centrifugation studies in normal and SLS cul-
tured fibroblasts. Cultured fibroblasts were fractionated as described
in Methods, and FALDH activity was assayed in each fraction. Data
are expressed as the enzyme specific activities measured in SLS cells
as a percentage ofthe mean activities in normal fibroblasts. Results
shown are the mean±SEM for five experiments using four normal
cell lines and three SLS cell lines.
Figure 5. Oxidation of free fatty aldehyde (octadecanal) and fatty
alcohol (octadecanol) to fatty acid by intact fibroblasts. Data represent
the results of two experiments and are the mean±SD using three
normal and four SLS cell lines. Open bars, normal controls; cross-
hatched bars, SLS.
Our results indicate that the primary defect in SLS involves the
FALDH component of FAO, which leads to impaired oxida-
tion of fatty aldehyde derived from fatty alcohol metabolism.
This FALDH component was deficient in SLS patients from
seven unrelated kindreds, eachpatient presumably the result of
independent mutations. No patient was found to be deficient
in the FADH component of FAO. Thus, our SLS patients
showed no evidence ofbiochemical heterogeneity. The obser-
vation that SLS heterozygotes have a partial deficiency in
FALDH activity is consistent with this enzyme being the pri-
mary genetic defect in SLS.
FAO has not been purified and little is known about its
subunit structure. At least four aldehyde dehydrogenases that
act on propionaldehyde have been identified in human tissues
(11), and it is possible that multiple FALDH enzymes also
exist, perhaps with overlapping substrate specificities. One or
more ofthese may function as a component ofthe FAO com-
plex. Fatty aldehyde dehydrogenases that are active against do-
decanal and shorter chain substrates have been purified from
rat liver(12, 13) and rabbit intestine (14). The rabbitFALDH is
capable offunctioning in the complete oxidation ofdodecanol
to fatty acid when it is reconstituted with FADH (7). However,
enzyme activity using substrates longer than 12-carbons was
not reported, and it is unclear whether this same enzyme is
analogous to the one deficient in SLS. Nevertheless, a single
FALDH enzyme may function as a component of the FAO
complex and catalyze the oxidation of a variety of aliphatic
aldehydes. The finding that SLS fibroblasts showed consider-
able deficiency in FALDH activity using aldehydes from 6 to
18 carbons long is most consistent with a singleFALDH acting
as a component of FAO. The residual FALDH activity mea-
sured in crude homogenates and subcellular fractions of SLS
cells may reflect the contribution ofother aldehyde dehydroge-
nases that preferentially act on different aldehyde substrates or
free fatty aldehydes. The presence ofnormal cytosolicFALDH
activity but decreased particulate enzyme activity in SLS is
FattyAldehyde Dehydrogenase Deficiency in Sjogren-Larsson Syndrome
consistent with the existence ofmultiple enzymesthatreact.on
long-chain aldehydes. Cytosolic aldehyde dehydrogenases are
known to have broad substrate specificities (15).
FALDH is thought to be a microsomal enzyme in rat liver
(12, 13) and rabbit intestine (14). Although a small amount of
FALDH activity appeared to be cytosolic, our differential cen-
trifugation studies indicated that most FALDH activity is par-
ticulate in normal fibroblasts. Density gradient separations of
cellular organelles,will be necessary to determine the subcellu-
lar localization ofFALDH and FAQ.
Intact SLS fibroblasts oxidized free fatty aldehyde nor-
mally, which suggeststhe presence ofmore than one ftyalde-
hyde oxidizing enzyme. It is possible that another FALDH or a
long-chain aldehydeoxidase exists, which isprimarily responsi-
ble foroxidizing free fatty aldehydeand is not deficient in SLS.
The role of the FAQ complex in recycling fatty alcohol is
most apparentin SLS patientswho are deficient in this enzyme
activity and exhibit symptoms involvingthe skin and nervous
system. With FAQ deficiency, it is expected that long-chain
alcohols would accumulate in SLS patients. The finding that
FAQ and FALDH activities in SLS were more severely de-
creased as the substrate chain length increased from 12 to 18
carbons is consistent with the previous report (6) that octade-
canol and hexadecanol accumulate in plasma from SLS pa-
tients, but tetradecanol does not. Presumably, SLS patients pos-
sess sufficient residual FAQ activity (35% ofnormal)with tetra-
decanol to prevent accumulation of this alcohol, whereas
residual enzyme activity is decreased to <20% ofnormal with
longer chain alcohols. The FALDH deficiency againsthexanal
and octanal raises the possibility that SLS patients may accu-
mulate medium-chain fatty alcohols. Judge et al. (16) have
recently demonstrated decreased activity ofhexanol oxidation
in the epidermis and jejunal mucosa of SLS patients using a
histochemical staining method.
Qur finding that all SLS patientswere deficient in FALDH
and none were deficient in the FADH component of FAQ
raises the possibilitythat the SLS phenotyperesultspartlyfrom
fatty aldehyde accumulation. Under FAQ assay conditions,
SLS fibroblast homogenates, accumulated fatty aldehyde de-
rived from fatty alcohol, whereas free aldehyde accumulation
could not be demonstrated in intact SLS fibroblasts. In the
intact SLS cell which possessessome NADH (and NADPH), it
is possible that the fatty aldehyde is reduced back to fattyalco-
hol. Alternately, the fatty aldehyde may covalently interact
with proteins or other molecules to form an aldehyde deriva-
tive, or be further converted to other metabolites.
Thebiological consequencesofeitherlong-chainalcohol or
aldehyde accumulation are unknown. Alcohols less than 14-
carbons long are known anesthetic agents, whereas fatty alco-
holslongerthan 14-carbons have little or no anesthetic proper-
ties (17). Long-chain alcohols have been shown to partition
into artificial lipid bilayers (17, 18) andsynapticvesicles (19).If
fatty alcohol accumulates in the skin of SLS patients, it could
modify the epidermal water barrier, which is critically depen-
dent on the lipid composition ofthe stratum corneum (20),and
lead to increased transepidermal water loss and ichthyosis.
Fatty aldehydes would also be expected to partition into cell
membranes, where the chemically reactive aldehyde group
may form covalent bonds with a variety of adducts and inter-
fere with the action of membrane-bound enzymes or disturb
membrane function. Acetaldehyde, for example, has been
shown to react with lysine residues in tubulin protein and in-
hibit microtubule assembly in vitro (21).
We thank the following physiciansfor kindly providing cell lines from
SLS patients: Drs. S. Black, G. Holmgren, E. Chaves, J. Bonnefont, and
This research was supported by National Institutes ofHealth grant
1. Sjdgren, T., and T. Larsson. 1957. Oligophreniain combination with con-
genital ichthyosis and spastic disorders. Acta Psychiatr. NeuroL. Scand. 32(Suppl.
2. Theile, U. 1974. Sjogren-[.arsson syndrome. Oligophrenia-ichthyosis -di/
tetraplegia. Humangenetik. 22:91-118.
3. Jagell, S., K.-H. Gustavson, and G. Holmgren. 1981. Sj6gren-L~arsson syn-
drome in Sweden: a clinical, genetic and epidemiological study. Clin. Genet.
4. Chaves-Carballo, E. 1987. Sjilgren-1.arsson syndrome. In Neurocutaneous
Diseases. M. R. Gomez, editor. Butterworth Publishers, Boston.
5. Rizzo, W. B., A. L. Dammann, and D. A. Craft. 1988. Sjdgren-Larsson
syndrome: impaired fatty alcohol oxidation in cultured fibroblasts due to defi-
cient fatty alcohol:nicotinamide adenine dinucleotide oxidoreductase activity.J.
Clin. Invest. 81:738-744.
6. Rizzo, W. B., A. L. Dammann, D. A. Craft, S. H. Black, A. Henderson
Tilton, D. Africk, E. Chaves-Carballo, G. Holmgren,and S. Jagell. 1989. Sj6gren-
Larsson syndrome: inherited defect in thefattyalcohol cycle. J. Pediatr. I115:228-
7. Ichihara, K., E. Kusunose, Y. Noda, and M. Kusunose. 1986. Some proper-
ties ofthe fatty alcohol oxidation system and reconstitution ofmicrosomal oxida-
tion activity in intestinal mucosa. Biochim. Biophys. Acta. 878:412-418.
8. Rizzo, W. B., A. L. Dammann, D. A. Craft, and M. W. Phillips. 1987. Fatty
alcohol metabolism in cultured human fibroblasts. Evidence for a fatty alcohol
cycle. J. Blot. Chem. 262:17412-17419.
9. Ferrell, W. J., and K.-C. Yao. 1972. Reductive and oxidative synthesisof
saturated and unsaturated fatty aldehydes. J. LipidRes. 13:23-26.
10. Kitson, T. M. 1977. The disulfiram-ethanol reaction. A review. J. Stud.
11. Harada, S., D. P. Agarwal, and H. W. Goedde. 1980. Electrophoreticand
biochemical studies of human aldehyde dehydrogenase isozymes in various tis-
12. Nakayasu, H., K. Mihara, and R. Sato. 1978. Purification and properties
of a membrane-bound aldehyde dehydrogenase from rat liver microsomes. Eio-
chem. Biophys. Res. Commun. 83:697-703.
13. Mitchell, D. Y., and D. R. Petersen. 1989. Oxidation ofaldehydic prod-
ucts oflipid peroxidation by rat liver microsomal aldehyde dehydrogenase.Arch.
Biochem. Biophys. 269:11-17.
14. Ichihara, K., Y. Noda, C. Tanaka, and M. Kusunose. 1986. Purification of
aldehyde dehydrogenase reconstitutively active in fatty alcohol oxidation from
rabbit intestinal microsomes. Biochim. Biophys. Adta. 878:419-425.
15. Weiner, H. 1980. Aldehyde oxidizing enzymes. In Enzymatic Basis of
Detoxication. Vol. 1. W. B. Jakoby, editor. Academic Press, New York. 26 1-280.
16. Judge, M. R., B. D. Lake, V. V. Smith, G. T. N. Besley, and J. I. Harper.
1990. Depletion ofalcohol (hexanol) dehydrogenase activityin the epidermisand
jejunal mucosa in Sj6gren-Larsson syndrome. J. Invest. Dermatol. 95:632-634.
17. Pringle, M. J., K. B. Brown, and K. W. Miller. 1981. Can the lipidtheories
ofanesthesia account for the cutoffin anesthetic potency in homologousseries of
alcohols? MoL. Pharmacol. 19:49-55.
18. Franks, N. P., and W. R. Lieb. 1986. Partitioning oflong-chainalcohols
into lipid bilayers: implications for mechanisms ofgeneralanesthesia. Proc. Nail.
Acad. Sdi. USA. 83:5116-5120.
19. Miller, K. W., L. L. Firestone, J. K. Alifimoff, and P. Streicher. 1989.
Nonanesthetic alcohols dissolve in synaptic membranes without perturbingtheir
lipids. Proc. Nail. Acad. Sci. USA. 86:1084-1087.
20. Elias, P. M. 1983. Epidermal lipids, barrier function, and desquamation.
J. Invest. Derinatol. 80:44-49.
21. Tuma, D. J., R. B. Jennett, and M. F. Sorrell. 1987. The interaction of
acetaldehyde with tubulin. Ann. NY. Acad. Sdi. 492:277-286.
W B. Rizzo andD. A. Craft