Differential gene expression in periportal and perivenous
mouse hepatocytes
Albert Braeuning
1
, Carina Ittrich
2
, Christoph Ko
¨
hle
1
, Stephan Hailfinger
1
, Michael Bonin
3
,
Albrecht Buchmann
1
and Michael Schwarz
1
1 Institute of Pharmacology and Toxicology, Department of Toxicology, University of Tuebingen, Germany
2 Central Unit of Biostatistics, German Cancer Research Center, Heidelberg, Germany
3 Institute for Human Genetics, Microarray Facility, Tuebingen, Germany
Hepatocytes play a pivotal role in both the synthesis
and degradation of numerous endogenous biomole-
cules, thus maintaining metabolic homeostasis, as well
as in the conversion and detoxification of xenobiotic
compounds. Based on the location of the blood ves-
sels, the terminal branches of the portal and the hep-
atic (central) veins and on the direction of the blood
flow, hepatocytes of each liver lobule can be divided
into two subpopulations, an upstream ‘periportal’ and

Tel: +49 7071 29 77398
E-mail:
(Received 24 July 2006, revised 15 Septem-
ber 2006, accepted 18 September 2006)
doi:10.1111/j.1742-4658.2006.05503.x
Hepatocytes located in the periportal and perivenous zones of the liver
lobule show remarkable differences in the levels and activities of various
enzymes and other proteins. To analyze global gene expression patterns of
periportal and perivenous hepatocytes, enriched populations of the two cell
types were isolated by combined collagenase ⁄ digitonin perfusion from
mouse liver and used for microarray analysis. In total, 198 genes and
expressed sequences were identified that demonstrated a ‡ 2-fold difference
in expression between hepatocytes from the two different zones of the liver.
A subset of 20 genes was additionally analyzed by real-time RT-PCR, val-
idating the results obtained by the microarray analysis. Several of the
differentially expressed genes encoded key enzymes of intermediary meta-
bolism, including those involved in glycolysis and gluconeogenesis, fatty
acid degradation, cholesterol and bile acid metabolism, amino acid degra-
dation and ammonia utilization. In addition, several enzymes of phase I
and phase II of xenobiotic metabolism were differentially expressed in peri-
portal and perivenous hepatocytes. Our results confirm previous findings
on metabolic zonation in liver, and extend our knowledge of the regulatory
mechanisms at the transcriptional level.
Abbreviations
GS, glutamine synthetase.
FEBS Journal 273 (2006) 5051–5061 ª 2006 The Authors Journal compilation ª 2006 FEBS 5051
subpopulation [4]. The list of zonally expressed
enzymes can be further extended to metabolic path-
ways such as glycogen synthesis, lipid metabolism and
bile acid formation [1]. Many enzymes of xenobiotic

regulation in the two hepatocyte subpopulations.
Results
Expression profiles of perivenous and periportal
hepatocytes
Perivenous and periportal mouse hepatocyte fractions
were obtained by combined digitonin ⁄ collagenase per-
fusion of liver of male C3H ⁄ He mice. The efficiency of
hepatocyte separation was monitored by western ana-
lysis of marker proteins with well-known zonal differ-
ences in expression, e.g. GS, a perivenous marker, and
E-cadherin, a periportal marker; this demonstrated the
expected differences in levels between periportal and
perivenous hepatocytes (Fig. 1). The RNA expression
patterns of hepatocyte fractions were analyzed by use
of the Affymetrix GeneChip MOE-430A, which con-
tains approximately 22 600 probe sets, including more
than 14 000 well-characterized mouse genes. Genes
were stratified into two groups according to their pref-
erential perivenous or preferential periportal expres-
sion, using as discriminators a Œlog
2
expression ratio Œ
‡ 1 (corresponding to a ‡ 2-fold difference in expres-
sion) and an adjusted P-value £ 0.1.
In total, we identified 198 genes and expressed
sequences that were differentially regulated in hepato-
cytes from the two different zones of the liver; 99 of
these were predominantly expressed in perivenous cells,
whereas another 99 were mainly expressed in peripor-
tal hepatocytes. A detailed list of the differentially

show positive log
2
ratios in the plot, whereas genes
with preferential periportal expression show negative
log
2
ratios.
To validate these data, the expression of 20 (10%)
of the zonated genes found in the microarray experi-
ment was additionally analyzed by real-time RT-PCR.
The PCR data closely resembled the findings of the
microarray analysis. A comparison of the results
obtained with the two methods is shown in Table 1.
Differences in genes encoding enzymes of
intermediary metabolism
The results of our present microarray analysis clearly
demonstrate differences between perivenous and peri-
portal hepatocytes in the expression of genes encoding
key enzymes of zonated pathways of intermediary meta-
bolism. This holds particularly true for genes encoding
enzymes playing a role in pathways that are known to
be stably zonated within the liver lobule. A schematic
representation of the observed differences in selected
metabolic pathways is given in Figs 3 and 4.
Glycolysis and gluconeogenesis
As shown in Fig. 3A, several genes encoding enzymes
participating in glycolysis are preferentially expressed in
perivenous cells. These include the genes encoding sorbi-
tol dehydrogenase (EC 1.1.1.14), aldehyde reductase
(EC 1.1.1.21), 6-phosphofructo-2-kinase (EC 2.7.1.105),

2
expression ratio| ‡ 1)
are indicated by horizontal and vertical lines;
these were chosen to identify genes with
significant alterations in expression (areas
indicated by gray). One hundred and
twenty-nine probe sets (corresponding to 99
transcripts) were predominantly expressed in
perivenous cells, showing positive log
2
ratios,
whereas another 114 probe sets (corres-
ponding to another 99 transcripts) were
mainly expressed in periportal hepatocytes,
showing negative log
2
ratios.
A. Braeuning et al. Zonal gene expression in mouse liver
FEBS Journal 273 (2006) 5051–5061 ª 2006 The Authors Journal compilation ª 2006 FEBS 5053
and cholesterol metabolism are shown in Fig. 3B. For
example, mRNAs coding for phosphatide phosphatase
(EC 3.1.3.4) and apoliprotein C-II, an essential co-
factor for the activation of lipoprotein lipase
(EC 3.1.1.34), are preferentially expressed in periportal
hepatocytes. As mentioned above, preferential peripor-
tal expression is also observed for the acetyl-CoA-
forming enzyme ATP citrate lyase (EC 2.3.3.8), which
provides acetyl-CoA for cholesterol synthesis. How-
ever, the mRNAs for HMG-CoA synthase and HMG-
CoA reductase, the enzymes catalyzing the initial steps

serine dehydratase (EC 4.3.1.17), and serine dehydra-
tase-like (EC 4.3.1.19). Oxaloacetate, the product of the
reaction catalyzed by the latter enzymes, can be used for
gluconeogenesis, a pathway that is also mainly located
in periportal hepatocytes [1].
Ammonia utilization
As shown in Fig. 3D, ammonia is used in perivenous
hepatocytes for glutamine synthesis, as GS
(EC 6.3.1.2), the key enzyme, is specifically expressed
in this hepatocyte subpopulation. Comparable zona-
tion is found for transporters participating in ammonia
(rhesus blood group-associated B glycoprotein) and
glutamate uptake (solute carriers 1A2 and 1A4), thus
providing the substrates for GS. In contrast, periportal
hepatocytes are lacking GS and use ammonia for urea
synthesis. With less stringent cutoff conditions, the
mRNAs of four enzymes of the urea cycle were found
to be preferentially localized in the periportal area,
namely ornithine transcarbamylase (EC 2.1.3.3), argin-
inosuccinate synthetase (EC 6.3.4.5), argininosuccinate
lyase (EC 4.3.2.1), and arginase (EC 3.5.3.1), showing
log
2
expression ratios between 0.58 and 0.82.
Xenobiotic metabolism
As expected, many genes encoding enzymes of xeno-
biotic metabolism were mainly expressed in perivenous
Table 1. Validation of microarray analysis data by real-time RT-PCR.
If genes are represented by more than one probe set on the chip,
their individual log

1.41 1.29
Phosphoenolpyruvate carboxykinase 1,
cytosolic (Pck1)
) 2.63 ) 3.12
Cytochrome P450 1a2 (Cyp1a2) 2.53 3.04
Cytochrome P450 2e1 (Cyp2e1) 1.69 4.12
Cytochrome P450 2f2 (Cyp2f2) ) 2.67 ) 4.58
Sulfotransferase 5a1 (Sult5a1) ) 3.03 ) 3.39
Aldehyde dehydrogenase 1B1
(Aldh1b1)
) 6.07 ) 4.98
Cytochrome P450 7a1 (Cyp7a1) 3.42 ⁄ 2.55 2.69
ATP citrate lyase (Acly) ) 1.36 ⁄ ) 1.44 ⁄
) 1.63
) 1.83
Cathepsin C (Ctsc) ) 3.77 ) 3.17
G protein-coupled receptor 49 (Gpr49) 2.63 8.10
Constitutive androstane receptor
(Nr1i3)
1.65 1.39
Aryl-hydrocarbon receptor (Ahr) 1.53 1.19
Hairy and enhancer of split 1 (Hes1) ) 1.90
) 1.49
Catenin beta interacting
protein 1 (Ctnnbip1)
) 1.81 ) 3.48
Cadherin 1 (Cdh1) ) 4.39 ) 5.58
Rhesus blood group-associated
B glycoprotein (Rhbg)
4.12 4.86

4.1.1.32
allosteric
activation
+
2.7.1.2
fructose-
1,6-bis-P
2.7.1.11
oxaloacetate
glucose-6-P
2.7.1.40
citrate
2.3.3.8
oxaloacetate citrate
A
1.1.1.41
glutamine
ammoniaglutamate
amino acid
degradation
Rhbg
Slc1A4Slc1A2
6.3.1.2
urea
cycle
2.6.1.13
ornithine
ammonia
D
urocanate

gluconeo-
genesis
diacylglycerol
diacylglycerol-P
3.1.3.4
APOC2
activation
+
3.1.1.34
B
Lipo-
proteins
Fig. 3. Zonal differences in expression of genes encoding enzymes and other proteins involved in intermediary metabolism. Perivenous
expression is indicated by green, and genes with preferential periportal expression are indicated by red. (A) Perivenous zonation of glycolysis
and periportal zonation of gluconeogenesis. (B) Fatty acid degradation and cholesterol metabolism in periportal hepatocytes. (C) Elevated
amino acid degradation in periportal hepatocytes. (D) Ammonia utilization for glutamine synthesis in perivenous hepatocytes. EC 1.1.1.14,
L-iditol-2-dehydrogenase (sorbitol dehydrogenase) (gene name: Sdh); EC 1.1.1.21, aldehyde reductase (Akr1b3); EC 2.7.1.2, glucokinase;
EC 2.7.1.105 ⁄ EC 3.1.3.46, 6-phosphofructo-2-kinase ⁄ fructose-2,6-bisphosphatase (bifunctional enzyme) (PfkFB1); EC 2.7.1.11, 6-phospho-
fructokinase; EC 2.7.1.40, pyruvate kinase liver and red blood cell (Pklr); EC 2.3.1.12, dihydrolipoamide-S-transferase (E2 component of pyru-
vate dehydrogenase complex) (Dlat ); EC 1.1.1.41, isocitrate dehydrogenase NAD
+
(Idh3a); EC 4.1.1.32, phosphoenolpyruvate carboxykinase
1, cytosolic (pck1); EC 2.3.3.8, ATP citrate lyase (Acly); APOC2, apolipoprotein C-II (essential cofactor for the activation of lipoprotein lipase);
EC 3.1.1.34, lipoprotein lipase; EC 3.1.3.4, phosphatide phosphatase type 2c (Ppap2c); EC 1.14.13.17, cytochrome P450 7A1 (cholesterol-7-
a-monooxygenase) (Cyp7a1); EC 4.3.1.3, histidine ammonia lyase (Hal); EC 4.2.1.49, urocanase domain containing 1 (urocanate hydratase)
(Uroc1); EC 2.1.2.5, glutamate formiminotransferase (Ftcd); EC 2.1.1.8, histamine-N-methyltransferase (Hnmt); EC 1.4.4.2, glycine decarboxy-
lase (part of glycine dehydrogenase complex) (Gldc); EC 4.3.1.17, serine dehydratase (Sds); EC 4.3.1.19, serine dehydratase-like (Sdsl);
EC 2.6.1.13, ornithine aminotransferase (oat); EC 6.3.1.2, glutamate ammonia ligase (glutamine synthetase) (glul); Slc1A2, solute carrier 1A2;
Slc1A4, solute carrier 1A4; Rhbg, rhesus blood group-associated B glycoprotein.
A. Braeuning et al. Zonal gene expression in mouse liver

observations on the zonation of glycolysis and glucone-
ogenesis in the liver, but also reveal zonal expression of
genes involved in glucose metabolism that have not
previously been reported as zonated. For example, peri-
venous localization was demonstrated for the mRNAs
of sorbitol dehydrogenase and aldehyde reductase,
which are involved in carbohydrate conversion
processes that provide glucose for further metabolism
in glycolysis. Neither of these enzymes has been pre-
viously reported to be differentially expressed between
perivenous and periportal hepatocytes. However, the
oxygen tension
glycolysis
gluconeogenesis
cholesterol
biosynthesis
bile acid synthesis
amino acid degradation
glutamine synthesis
metabolism of
xenobiotics
blood flow
fatty acid degradation
portal vein central vein
O
2
O
2
hormones,
growth factors

in perivenous hepatocytes is in accordance with
the known perivenous activity of this enzyme [11,12].
The only gene of the glycolytic pathway that is mainly
expressed in periportal hepatocytes is that encoding
pyruvate kinase. The metabolic capacity of the res-
pective protein, however, was found to be equally
distributed throughout the liver lobuli [13], or to be
even higher in the perivenous zone [14], suggesting a
post-transcriptional regulation mechanism for this
enzyme. On the other hand, periportal zonation
of gluconeogenesis and particularly of phosphoenol-
pyruvate carboxykinase, has been reported before [1].
Fatty acid degradation is another metabolic pathway
underlying zonal expression in liver, as two genes,
those encoding phosphatide phosphatase and apolipo-
protein C2, a cofactor for activation of lipoprotein
lipase, were found in our study to be preferentially
expressed in the periportal hepatocyte subpopulation.
Zonal-specific expression of these genes has not been
reported so far, but our findings are in line with previ-
ous observations of the periportal localization of fatty
acid degradation [15]. Bile acid synthesis from choles-
terol takes place in perivenous hepatocytes, as the key
enzyme of this metabolic pathway, cholesterol-7-a-
monooxygenase, is preferentially expressed in hepato-
cytes surrounding the central veins [16]. Our analysis
confirms previous findings on the zonation of choles-
terol-7-a-monooxygenase protein [17] and mRNA [16].
The mRNA levels for ATP citrate lyase, an enzyme
forming acetyl-CoA from citrate, thus providing sub-

2
ratios.
Gene
Log
2
ratio(s)
pv versus pp
Cytochrome P450 2a4 ⁄ 2a5 6.43
Glutathione-S-transferase mu 3 4.36 ⁄ 2.31
Glutathione-S-transferase mu 2 4.19
Carboxylesterase 2 3.59
Glutathione-S-transferase mu 6 3.13
Cytochrome P450 2c50 ⁄ 2c54 2.98
Cytochrome P450 1a2 2.53
Cytochrome P450 2c55 2.35
Cytochrome P450 2c29 2.34
Cytochrome P450 2g1 2.00
Cytochrome P450 oxidoreductase 1.85
Cytochrome P450 2e1 1.69
Constitutive androstane receptor 1.65
Aryl-hydrocarbon receptor 1.53
Cytochrome P450 2c38 1.48
Glutathione-S-transferase alpha 3 1.34
Sulfotransferase 1B1 1.16
Sulfotransferase 1D1 1.03 ⁄ 1.01
Glutathione-S-transferase mu 1 1.01
Glutathione-S-transferase alpha 2 ) 1.09
Arsenic methyltransferase ) 1.51
Cytochrome P450 2f2 ) 2.67
Sulfotransferase 5a1 ) 3.03

been reported to be mainly localized in periportal
hepatocytes [1,2]. Our data now suggest that this
localization is based on differences in mRNA levels of
four enzymes of the urea cycle.
Whereas the zonal-specific expression of the main
enzymes of drug and xenobiotic metabolism has been
the subject of extensive research (e.g. cytochrome P450
zonation [5]), the zonal expression profiles of the more
uncommon cytochrome P450 isoforms have mostly not
been described in the literature. For example, up to
now cytochrome P450 2G1 (Cyp2g1) has been consid-
ered to be exclusively expressed in the olfactory
mucosa in mammals [24]. Whereas the mRNAs for
most xenobiotic-metabolizing enzymes are mainly
expressed in hepatocytes near the central veins, a small
number of these enzymes exhibit preferential periportal
expression. The periportal localization of these
mRNAs has not been reported in the literature so far.
The mechanisms underlying zonal gene expression in
the liver are not yet fully understood. Based on compari-
sons of mRNA ⁄ protein expression patterns of perive-
nous and periportal hepatocytes with those of liver
tumors containing activating mutations in either the
Ha-ras or ctnnb1 (catnb; b-catenin) gene, we developed
the hypothesis that two opposing signaling pathways
triggered by Ha-ras- and b-catenin-dependent factors
may determine zonal differences in gene expression in
murine liver [22]. In addition, the adenomatous poly-
posis coli (APC) tumor suppressor gene, an important
regulator of b-catenin signaling, has also been estab-

were isolated and enriched by combined digitonin ⁄ collage-
nase perfusion of the liver according to Taniai et al. [26],
with minor modifications as described previously [22]. First,
the liver was perfused for 10 min with Krebs ⁄ Henseleit buf-
fer at 37 °C. To obtain periportal hepatocyte subpopula-
tions, a 5 mm digitonin solution was infused for 10 s
through the vena cava and then immediately flushed out
from the opposite direction. To obtain perivenous hepato-
cytes, the digitonin solution was infused through the portal
vein. After digitonin treatment, the liver was perfused with
collagenase solution. Subsequently, viable hepatocytes were
separated by density gradient centrifugation. Viability of
the resulting hepatocyte fractions was always  80–90% as
determined by trypan blue staining. The efficiency of
Zonal gene expression in mouse liver A. Braeuning et al.
5058 FEBS Journal 273 (2006) 5051–5061 ª 2006 The Authors Journal compilation ª 2006 FEBS
separation of hepatocytes into periportal and perivenous
subfractions was determined by real-time RT-PCR analysis
of GS expression and western blotting for marker proteins
as described below.
Microarray analysis and statistical evaluation
of data
The Affymetrix GeneChip MOE-430A (Affymetrix, Santa
Clara, CA, USA) was used for mRNA expression profiling.
Six chips were hybridized with cRNA from three periportal
and three perivenous hepatocyte isolates, obtained from
independent liver perfusions. RNA quality was controlled
with the Laboratory-on-Chip-System Bioanalyzer 2100
(Agilent, Palo Alto, CA, USA). Data normalization and
statistical analysis was carried out essentially as previously

3
VO
4
] buffer plus protease inhibitor cocktail
(Complete Mini, Roche, Mannheim, Germany). Protein
concentrations were estimated using the Bradford assay.
Western blotting was carried out as recently described [22]
using antibodies against GS (1 : 5000 dilution; Sigma, Tauf-
kirchen, Germany), E-cadherin (1 : 1000; Transduction
Laboratories, Lexington, KY, USA), G-protein-coupled
receptor 49 (1 : 1000; Affinity BioReagents, Golden, CO,
USA), glyceraldehyde-3-phosphate dehydrogenase (1 : 1,000;
Chemicon, Hampshire, Chandler’s Ford, UK) and cyto-
chrome P450 1A (1 : 1000; gift of R Wolf, Biomedical
Research Centre, University of Dundee, UK). Antibody
binding was visualized using appropriate alkaline phospha-
tase-conjugated secondary antibodies (1 : 10 000; Tropix,
Applied Biosystems, Weiterstadt, Germany) and CDP-Star
as a substrate. Chemoluminescence signals were monitored
by use of a CCD camera system.
Quantitative determination of mRNAs by RT-PCR
Total RNA was isolated with Trizol reagent (Invitrogen,
Karlsruhe, Germany). RNA was purified using the RNeasy
Table 3. PCR primers.
Gene
Forward
(5¢-to3¢)
Reverse
(5¢-to3¢)
Glul GCGAAGACTTTGGGGTGATA GTGCCTCTTGCTCAGTTTGTC

primers. Expression analysis was performed
using the LightCycler real-time PCR system (Roche,
Mannheim, Germany). Expression of 18S rRNA was used
for normalization. The primer pairs used for PCR amplifi-
cation are given in Table 3.
Acknowledgements
We gratefully acknowledge the excellent technical
assistance of Elke Zabinsky and Silvia Vetter. We also
thank Dr R. Wolf for the gift of Cyp1A antibody.
This study was supported by the Deutsche Krebshilfe
(grant 106356).
References
1 Jungermann K & Katz N (1989) Functional speciali-
zation of different hepatocyte populations. Physiol Rev
69, 708–764.
2 Gebhardt R (1992) Metabolic zonation of the liver:
regulation and implications for liver function. Pharma-
col Ther 53, 275–354.
3 Jungermann K & Kietzmann T (1996) Zonation of par-
enchymal and nonparenchymal metabolism in liver.
Annu Rev Nutr 16, 179–203.
4 Gebhardt R & Mecke D (1983) Heterogeneous distribu-
tion of glutamine synthetase among rat liver parenchymal
cells in situ and in primary culture. EMBO J 2, 567–570.
5 Oinonen T & Lindros KO (1998) Zonation of hepatic
cytochrome P-450 expression and regulation. Biochem J
329, 17–35.
6 Lemaigre FP & Rousseau GG (1994) Transcriptional
control of genes that regulate glycolysis and gluconeo-
genesis in adult liver. Biochem J 303, 1–14.

origin. Arch Biochem Biophys 326 , 202–206.
14 Zierz S, Katz N & Jungermann K (1983) Distribution
of pyruvate kinase type L and M2 in microdissected
periportal and perivenous rat liver tissue with different
dietary states. Hoppe Seylers Z Physiol Chem 364,
1447–1453.
15 Guzman M & Castro J (1989) Zonation of fatty acid
metabolism in rat liver. Biochem J 264, 107–113.
16 Twisk J, Hoekman MF, Mager WH, Moorman AF, de
Boer PA, Scheja L, Princen HM & Gebhardt R (1995)
Heterogeneous expression of cholesterol 7 alpha-hydro-
xylase and sterol 27-hydroxylase genes in the rat liver
lobulus. J Clin Invest 95, 1235–1243.
17 Ugele B, Kempen HJ, Kempen JM, Gebhardt R, Meijer
P, Burger HJ & Princen HM (1991) Heterogeneity of
rat liver parenchyma in cholesterol 7 alpha-hydroxylase
and bile acid synthesis. Biochem J 276, 73–77.
18 Evans JL, Quistorff B & Witters LA (1989) Zonation of
hepatic lipogenic enzymes identified by dual-digitonin-
pulse perfusion. Biochem J 259, 821–829.
19 Katz NR, Fischer W & Ick M (1983) Heterogeneous
distribution of ATP citrate lyase in rat-liver parench-
yma. Microradiochemical determination in microdis-
sected periportal and perivenous liver tissue. Eur J
Biochem
130, 297–301.
20 Ogawa H & Kawamata S (1995) Periportal expression
of the serine dehydratase gene in rat liver. Histochem J
27, 380–387.
21 Weiner ID, Miller RT & Verlander JW (2003) Localiza-

A et al. (2005) Genotype–phenotype relationships in
hepatocellular tumors from mice and man. Hepatology
42, 353–361.
Supplementary material
The following supplementary material is available
online:
Table S1. Gene expression in perivenous versus peri-
portal hepatocytes.
This material is available as part of the online article
from
A. Braeuning et al. Zonal gene expression in mouse liver
FEBS Journal 273 (2006) 5051–5061 ª 2006 The Authors Journal compilation ª 2006 FEBS 5061