Báo cáo khoa học: Differential expression of liver and kidney proteins in a mouse model for primary hyperoxaluria type I pdf

Homozygous Agxt ⁄ mice show severe hyperoxaluria, Keywords hyperoxaluria; kidney; liver; mouse model; subcellular fractions Correspondence E.. To improve our understanding of the metabol

mouse model for primary hyperoxaluria type I

Juan R Herna´ndez-Fernaud1 and Eduardo Salido2

1 Max Planck Institute for Biochemistry, Department of Proteomics and Signal Transduction, Klopferspitz, Martinsried, Germany

2 Hospital Universitario Canarias, Center Biomedical Research on Rare Diseases (CIBERER) and Institute of Biomedical Technologies (ITB), Tenerife, Spain

Introduction

Primary hyperoxaluria type I (PHI) is a rare autosomal

recessive disease caused by mutations in the

alanine-gly-oxylate aminotransferase gene (AGXT)

Alanine-glyoxy-late aminotransferase (AGT) (or alanine-glyoxyAlanine-glyoxy-late

aminotransferase 1, AGT1), the protein encoded by

AGXT, plays an important physiological role in

glyoxy-late detoxification by converting it into glycine The

enzyme is present in peroxisomes and⁄ or mitochondria

in different mammalian species, with peroxisomal AGT

being mainly responsible for the detoxification of

glyco-late-derived glyoxylate, and mitochondrial AGT playing

a major role in the metabolism of

hydroxyproline-derived glyoxylate [1] In humans, insufficient AGT

activity in peroxisomes leads to increased cytosolic

conversion of glyoxylate to oxalate Excessive renal excretion of oxalate causes calcium oxalate deposition (nephrocalcinosis and urolithiasis) and eventual loss of renal function After renal failure, calcium oxalate depo-sition becomes widespread and life-threatening unless liver and kidney transplantation are performed With a better understanding of glyoxylate metabolism, sub-strate depletion may potentially be a useful intervention

in patients with PHI [2]

In order to further analyze the mechanisms of PHI disease, and to explore new therapeutic approaches, we have developed an Agxt knockout (AgxtKO) mouse that reproduces some key features of PHI [3] Homozygous Agxt) ⁄ )mice show severe hyperoxaluria,

Keywords

hyperoxaluria; kidney; liver; mouse model;

subcellular fractions

Correspondence

E C Salido, Hospital Universitario Canarias,

Center Biomedical Research on Rare

Diseases (CIBERER) and Institute of

Biomedical Technologies (ITB), Tenerife

38320, Spain

Fax: +34 922 647 112

Tel: +34 922 319 338

E-mail: esalido@ull.es

(Received 29 July 2010, revised 3

September 2010, accepted 10 September

2010)

doi:10.1111/j.1742-4658.2010.07882.x

Mutations in the alanine-glyoxylate aminotransferase gene (AGXT) are responsible for primary hyperoxaluria type I, a rare disease characterized

by excessive hepatic oxalate production that leads to renal failure A deeper understanding of the changes in the metabolic pathways secondary to the lack of AGXT expression is needed in order to explore substrate depletion

as a therapeutic strategy to limit oxalate production in primary hyperoxal-uria type I We have developed an Agxt knockout (AgxtKO) mouse that reproduces some key features of primary hyperoxaluria type I To improve our understanding of the metabolic adjustments subsequent to AGXT defi-ciency, we performed a proteomic analysis of the changes in expression lev-els of various subcellular fractions of liver and kidney metabolism linked

to the lack of AGXT In this article, we report specific changes in the liver and kidney proteome of AgxtKO mice that point to significant variations

in gluconeogenesis, glycolysis and fatty acid pathways

Abbreviations

AGT1, alanine-glyoxylate aminotransferase 1; AGT2, alanine-glyoxylate aminotransferase 2; AGXT, alanine-glyoxylate aminotransferase gene; Agxt) ⁄ ), alanine-glyoxylate aminotransferase homozygous knockout; ML, mitochondrial ⁄ lysosomal; PHI, primary hyperoxaluria type I; SPT, serine-pyruvate aminotransferase; 2-DE, two-dimensional electrophoresis.

and males develop calcium oxalate crystalluria and

cal-culi in the urine bladder, although no deposits in the

renal parenchyma (nephrocalcinosis) are observed

unless the animals are subjected to metabolic overload

To better characterize this model, and provide

evidence useful in substrate depletion strategies, we

report in this article a proteomic analysis of the

changes in expression levels of various enzymes of liver

and kidney metabolism linked to the lack of AGT

Results

We first attempted to detect differences in protein

expression between hyperoxaluric and control mice at

the whole-organ proteome level, using either liver

or kidney samples This approach yielded insufficient

protein spots and lower reproducibility than that based

on subcellular fractionation, and was not pursued fur-ther However, for each subcellular fraction studied, more than 300 protein spots were detected in each two-dimensional electrophoresis (2-DE) gel For each frac-tion, three 2-DE silver- and Coomassie-stained gels were integrated and analyzed, and high reproducibility was achieved (Fig 1) By image analysis, using the relative spot volume parameter, the comparison between gels of wild-type and knockout kidney proteomes revealed 22 spots whose protein levels were significantly different between groups (P < 0.01), with three exclusive of knockout mice Twenty of these differentially expressed proteins were correctly matched to protein candidates

in the database (Table 1) according to their peptide mass fingerprints analyzed by MALDI-TOF MS

Fig 1 Comparison of 2-DE patterns among

different extraction methods and cell

fractions (A, B) Total extraction protein

method of kidney and liver organs,

respectively (C, D, E)

Mitochondrial-lysosomal, peroxisomal and cytosolic

fractions of kidney (F, G, H)

Mitochondrial-lysosomal, peroxisomal and cytosolic

fractions of liver Total protein (300 lg) was

subjected to 2-DE (first dimension: glass

capillaries; pH 3–10; 12 cm; second

) Proteins were visualized by

silver staining.

Database search and functional exploration of these

proteins revealed that they were associated with

dif-ferent metabolic aspects, such as oxidoreductase

activ-ity, glycolysis, glycine, glyoxylate, fatty acid and

pyruvate metabolism Hydroxyacid oxidase 3 was

two-fold more abundant in knockout mice than in

controls In contrast, d-amino acid oxidase 1 was

2.3-fold downregulated in hyperoxaluric mice Enolase 1

and malic enzyme were upregulated Furthermore,

acyl-coenzyme A dehydrogenase, mercaptopyruvate

sulfotransferase and abhydrolase domain protein were

only detected in knockout mouse kidneys (Table 1,

Fig 2A)

In liver fractions, 18 spots were identified with protein

levels significantly different between the groups

(P < 0.01), and two were exclusively detected in

knock-out mice In 14 of the 18 spots, MALDI peptide mass

fingerprints allowed the identification of the

correspond-ing proteins in the database (Table 2) Database search

and functional exploration of these proteins revealed

that they were associated with gluconeogenesis and

glycolysis In this sense, fructose bisphosphatase was

2.4-fold upregulated in knockout mice However,

alde-hyde dehydrogenase, carbonic anhydrase, enolase and

malic enzyme were downregulated (Table 2, Fig 2B) In

cytosolic fractions, the fumarylacetoacetate hydrolase

and peroxiredoxin 6 appeared, with shifted pI from

approximately 6.9 to 7 and 6 to 5.5, respectively

Western blot analysis was used to confirm the main

differences in expression found in 2-DE gels, provided

that antibodies were available

The results are summarized in Fig 3A In AgxtKO

mice, kidney enolase was clearly overexpressed, as were

liver fructose bisphosphatase and catalase, whereas

liver enolase and carbonic anhydrase 3 were

downregu-lated Comparable amounts of b-actin were present in

AgxtKO and wild-type cytosolic fractions, and the

absence of AGT1 protein in AgxtKO samples was also

confirmed by western blot

The changes in expression levels observed in these

few proteins are in agreement with 2-DE results, which

is consistent with the reliability of our comparative

proteomic study

To assess the tissue specificity of the liver and

kid-ney response, we also performed western blot analysis

of skeletal muscle proteins We observed high

variabil-ity and could not reproduce the detected differences in

liver and kidney samples (Fig 3B)

Discussion

We have analyzed the changes in protein expression

within the liver and kidney of Agxt) ⁄ ) deficient mice

compared with wild-type controls by 2-DE separation and MS The analysis of specific subcellular fractions was necessary to obtain highly informative and repro-ducible 2-DE gels The modified fractionation protocol adopted has been used previously in proteomic studies [4], but does not result in highly pure fractions, which

is likely to be the reason for some inconsistencies between the fraction in which we detected a differen-tially expressed protein and their accepted subcellular localization For instance, we detected d-amino acid oxidase in the mitochondrial⁄ lysosomal (ML) fraction

of kidney, whereas its accepted localization is either cytosolic or peroxisomal Most likely, our ML fraction contained peroxisomes that cosedimented during the procedure used Similarly, liver catalase was detected

in our cytosolic fraction, indicating that peroxisomes and⁄ or peroxisomal proteins were still present in the supernatant after the 7300 g centrifugation Under standard purification procedures, peroxisomal proteins are known to contaminate other subcellular fractions because of peroxisomal fragility With this limitation, our fractionation method was mainly useful as a sim-ple way to reduce the comsim-plexity of the proteome, facilitating the differential expression analysis between wild-type and AgxtKO mice

Agxt) ⁄ ) mice have impaired glyoxylate detoxifica-tion, with subsequent oxalate overproduction by the liver and increased urinary oxalate excretion, similar to patients with PHI [3] However, significant differences between mouse and human glyoxylate and glucose metabolism must be considered Although human AGT1, the product of the AGXT gene, is predomi-nantly localized in the peroxisome, the mouse Agxt1 gene is transcribed into two different mRNA species, coding for mitochondrial and peroxisomal variants [5] Indeed, rodent AGT1 is also known as serine-pyruvate aminotransferase (SPT) because the mitochondrial form participates in gluconeogenesis from serine, whereas the conversion of glyoxylate to glycine takes place largely in peroxisomes No alterations of glucose metabolism have been described in patients with PHI

In AgxtKO mice, we detected an increase in liver fructose-1,6-bisphosphatase, an enzyme involved in the hydrolysis of fructose-1,6-bisphosphate, which plays an important regulatory role in gluconeogenesis [6] In the same hepatic fractions, a decrease in cytosolic malic enzyme 1 was observed, pointing to a reduction in NADPH available for fatty acid biosynthesis Taken together, these results seem to be indicative of an adap-tation in favor of liver gluconeogenesis in response to the lack of AGT1 Other downregulated enzymes, such as aldehyde dehydrogenase 2, enolase 1, UDP-glucose pyrophosphorylase 2 and fumarylacetoacetate

Group number

Theorical MW

Matched peptide Sequence coverage

Mascot score Missed cleavage

a The

hydrolase, appear to support this observation These

results are consistent with our previous observation

that AgxtKO mice did not seem to show a deficit in

gluconeogenesis despite the absence of the AGXT1

gene product [3] There is also a significant level of

another aminotransferase, AGT2, in mouse liver [7],

although kinetic studies [8] indicate that its

alanine-glyoxylate aminotransferase activity is not favored

over aminobutyrate-pyruvate, b-alanine-pyruvate and

dimethylarginine-pyruvate aminotransferase activities

In the rat, gluconeogenesis from l-serine takes place

mainly through l-serine dehydratase, whereas the flux

through SPT⁄ AGT in gluconeogenesis from serine has

been shown to be significant only after the liver

mito-chondrial form of the AGT1 enzyme had been induced

by glucagon [9] However, the peroxisomal form of

SPT⁄ AGT predominates during constitutive expression

of rat and mouse AGXT genes, and the gluconeogenic flux from serine also takes place in this organelle to some extent [10] Amino acid metabolism is considered

to be a major contributor to endogenous oxalate syn-thesis, justifying the study of changes in liver enzymes

in the context of primary hyperoxaluria It could be speculated that our finding of enhanced liver gluconeo-genesis in the PHI mouse model is an adaptation to the lack of serine flux through AGT, and modifications that potentiate neoglucogenesis might be beneficial in primary hyperoxaluria, reducing the oxalate contribu-tion from amino acid metabolism These modificacontribu-tions might be seen as a form of substrate depletion How-ever, the above-mentioned differences in AGT subcel-lular localization between humans and laboratory

Fig 2 Metabolic kidney (A) and liver (B) enzymes upregulated (+) and downregulated

in knockout mice Acads, acyl-coenzyme A dehydrogenase, short chain; Aco1,

aminotransferase knockout; Aldh2, aldehyde dehydrogenase 2; Car3, carbonic

oxidase 1; Eno1, enolase 1, a non-neuron; Fah, fumarylacetoacetate hydrolase; Fbp1, fructose bisphosphatase 1; Hao3, hydroxyacid oxidase 3; Me1, malic enzyme 1, NADP(+)-dependent; Mpst, mercaptopyruvate sulfotransferase; Pgam1, phosphoglycerate mutase 1; Prdx6, peroxiredoxin 6; Ugp2, UDP-glucose pyrophosphorylase 2.

Group number

Theorical MW

Matched peptide Sequence coverage

Mascot score Missed cleavage

rodents pose a major limitation to our study and the

inferences that can be based on our findings Further

studies of the sources of endogenous oxalate synthesis

in humans are needed

The phenotypic features of Agxt) ⁄ )mice are

proba-bly the direct consequence of impaired glyoxylate

detoxification, with subsequent oxalate overproduction

by the liver and increased urinary oxalate excretion

As AGT1 is not expressed at significant levels in the

kidney, the changes observed in the kidney proteome

could be a consequence of variations in filtered

metab-olites, such as oxalate, present at high levels in

AgxtKO mice Nephrocalcinosis is essentially absent in

Agxt) ⁄ ) mice, despite high urinary oxalate excretion,

unless glyoxylate precursors are administered Thus,

the response in the kidney proteome to AGT1 deficit is

unlikely to be secondary to serious tissue damage The

increase in kidney enolase points to an enhanced

gly-colysis, whereas higher levels of hydroxyacid oxidase 3

could represent adjustments in medium-chain

hydroxy-fatty acid metabolism The overexpression of enolase 1

and malic enzyme 1 supports the induction of fatty

acid metabolism The reduction in d-amino acid

oxi-dase 1 expression in the kidney proteome is interesting,

in view of the contribution of this enzyme to

glyoxy-late production from glycine Thus, it could be

specu-lated that a decrease in d-amino acid oxidase 1

expression might be aimed at reducing the glyoxylate

overload in the kidney of AgxtKO mice

In conclusion, changes in the proteome contents of hyperoxaluric liver and kidney subcellular fractions should improve our understanding of the metabolic adjustments subsequent to AGXT deficiency, and might provide relevant clues for future developments

of substrate depletion approaches in the treatment of primary hyperoxaluria

Experimental procedures

Animals

Mice were bred and maintained in a pathogen-free facility, with free access to standard chow (A04, SAFE, Augy,

genotyped as reported previously [3] All the experiments were performed in accordance with Spanish and European law regarding the use of animals in research (European Community Council Directive of 24 November 1986,

by our Institutional Committee on Ethics in Animal Experi-mentation (CEBA-HUC) Mice were sacrificed by cervical dislocation, performed by trained personnel, in accordance with CEBA-HUC-approved protocols Immediately after brain death, laparotomy and bilateral thoracotomy were performed and the organs were harvested, making sure that the animals did not suffer at any stage of the procedure Tissues from eight, 3-month-old male mice allocated to

0

1

2

3

4

5

6

7

8

9

Eno1-K Fbp1-L

Eno1-M

Eno1-M FC

0

1

2

3

A

B

Fig 3 (A) Western blot analyses of enolase

hydroxy-pyruvate reductase (Grhpr-K) from kidney and fructose bisphosphatase (Fbp1-L), carbonic anhydrase 3 (Car3-L), catalase (Cat-L), enolase (Eno1-L), alanine-glyoxylate aminotransferase (Agt-L), actin (Actin-L) and

reductase (Grhpr-L) from liver (B) Western blot analyses of enolase (Eno1-M), fructose bisphosphatase (Fbp1-M), carbonic anhydr-ase 3 (Car3-M) and catalanhydr-ase (Cat-M) from muscle The analyses are in

cytosolic (FC) fractions Densitometry of western blot bands was performed and units of selected proteins were calculated.

A representative western blot band for each protein of each experimental group is shown Relative molecular weights were 47,

74, 29, 60, 45, 40 and 36 kDa for Eno1, Fbp1, Car3, Cat, Agt, Actin and Grhpr, respectively.

Agxt) ⁄ ) (hyperoxaluric) and from homozygous wild-type

thor-oughly rinsed in ice-cold saline before freezing

Sample preparation

Tissues for whole-protein extraction were frozen and

crushed in liquid nitrogen The powder was lyophilized and

extraction buffer with 8 m urea, 4% Chaps, 40 mm Tris,

65 mm 1,4-dithioerythritol, 0.05% SDS and 2%

ampho-lytes Next, the sample was centrifuged at 13 000 g for

Subcellular fractionation by differential centrifugation was

performed as described previously [4] The tissue was

contain-ing 250 mm sucrose, 10 mm Tris⁄ HCl, pH 7.5, and 1 mm

EDTA The cells were ruptured by 20 strokes in a glass

homogenizer, and the lysate was centrifuged at 200 g for

10 min to sediment the nuclei The supernatant was

centrifuged again at 2000 g for 10 min to sediment a

mitochondria-containing pellet that was homogenized in

1 mL of isotonic buffer and centrifuged at 7300 g for 10 min

to obtain a crude mitochondrial pellet The 2000 g

superna-tant was centrifuged at 7300 g for 10 min to obtain a crude

peroxisomal fraction The peroxisome pellet was

homoge-nized and centrifuged once again, and the supernatant was

centrifuged for 60 min at 7300 g to remove additional

organ-elles from the cytosolic fraction Proteins in the cytosolic

once and air dried Organelle and protein pellets were

solubilized in the extraction buffer and centrifuged at

determined by the Bradford method with BSA as the

pro-tein standard

2-DE

Isoelectric focusing was performed using glass capillary

tubes (inside diameter, 1.5 mm; length, 12 cm) For

separa-tion in the pH 5–8 range, capillary tubes were filled with

solution containing 3% acrylamide, 7 m urea, 0.6% Triton

X-100, 0.75% ampholytes pH 5–8, 0.22% ampholytes

N,N,N¢,N¢-tetramethylethylenediamine and 0.08%

ammo-nium persulfate For separation in the pH 3–10 range,

0.75% ampholytes pH 3–10 and 0.22% ampholytes pH 5–8

300 lg of total protein were applied to the basic end of the

tube gel Cathodic and anodic buffers were 20 mm NaOH

consisted of 1 h at 100 and 1 h at 300 V, followed by

17.5 h at 1000 V and 30 min at 2000 V Next, the capillar-ies were equilibrated for 15 min in reducing buffer

SDS and 1% dithiothreitol, followed by a blocking step in similar buffer containing 2.5% iodoacetamide instead of dithiothreitol for another 15 min The capillary gels were

0.5% low-melting agarose containing a trace of

for 15 min and then at 50 mA per gel until the blue front reached the bottom For external calibrations, molec-ular mass markers (Sigma, St Louis, MO, USA) were loaded onto the second dimension The protein spots were visualized by staining with either Coomassie blue R-250 for preparative gels [11] or silver nitrate for analyti-cal gels [12]

Image capture and analysis

Gels were scanned using a UMAX scanner (Amersham Biosciences, Barcelona, Spain) and the images were

Geneva, Switzerland), including spot detection, quantifica-tion, normalizaquantifica-tion, data analysis and statistics Compara-tive analysis of protein spots was performed by matching corresponding spots across different gels Each of the matched protein spots was rechecked manually Intensity volumes of individual spots were normalized with the total intensity volume of all spots present in each gel before per-forming differential expression analysis The Kolmogorov– Smirnov test was used to assess the statistical significance

of the differences between the normalized intensity volumes

expressed proteins were excised and subjected to subsequent identification by MS

MALDI-TOF MS

Protein spots were manually excised from Coomassie-stained gels, and tryptic in-gel digestion and desalting steps were performed using 96-well ZipPlates (Millipore, Bedford,

MA, USA) according to the manufacturer’s instructions The resulting peptides were mixed with 1 lL of

Anchor-chip plates as described by the manufacturer (Bruker Daltonics, Bremen, Germany) Peptide mass fingerprint spectra were measured on an Autoflex MALDI-TOF mass spectrometer (Bruker-Daltonics) in a positive ion reflection mode at an accelerating voltage of 20 kV, and spectra in the 900–3200 Da range were recorded For one main spec-trum, 30 subspectra with 30 shots per subspectrum were accumulated A pepmix calibration kit (Bruker-Daltonics) was used for calibration and the standard mass deviation

was < 10 ppm The peak lists were created with flex

fol-lows: SNAP peak detection algorithm; signal-to-noise ratio,

10; quality factor threshold, 30; maximal 100 peaks per

spot The peptide mass fingerprints were rechecked

manu-ally Peptide mass fingerprint data were submitted to the

MASCOT search engine for protein identification using the

Mascot database The search parameters were set according

to the following criteria: Mus musculus for taxonomy;

carb-amidomethyl (C) for fixed modifications; oxidation (M) for

variable modifications; and ±100 ppm for peptide ion mass

tolerance

Western blot

Subcellular fractions from six mice were obtained as

described above Protein concentration was measured using

the Bradford method, and 50 lg of protein were analyzed

by immunoblotting [13] with anti-AGT affinity-purified

rab-bit serum, anti-carbonic anhydrase 3 (1:1000 dilution) goat

serum or anti-fructose-1,6-bisphosphatase (1:1000) rabbit

(ABCAM, Cambridge, UK) and anti-catalase (1:5000

dilu-tion) mouse IgG1 (Sigma-Aldrich, St Louis, MO, USA)

Peroxidase-conjugated anti-rabbit IgG, anti-mouse IgG or

anti-goat IgG (Jackson Immunoresearch, West Grove, PA,

USA) was used as secondary antibody and the

chemilumi-nescence substrate was obtained from Pierce (Rockford, IL,

USA) Controls to ensure that equal amounts of protein

were loaded per lane involved Coomassie staining of

parallel gels, and reprobing the membranes with mouse

anti-b-actin (1 : 10 000; Sigma) and rabbit anti-glyoxylate

against recombinant protein) antibodies Inmunoreactive

specific bands were quantified using a UMAX scanner

(GeneBio) The signal was normalized against each

group to obtain the protein changes as expression

ratios

Acknowledgements

We are grateful to Cristina Paz for excellent technical

help The study was supported by grant 2007-62343

(Spanish Ministry of Science)

References

1 Behnam JT, Williams EL, Brink S, Rumsby G &

Danpure CJ (2006) Reconstruction of human

hepatocyte glyoxylate metabolic pathways in stably

transformed Chinese-hamster ovary cells Biochem J

394, 409–416

2 Danpure CJ (2006) Primary hyperoxaluria type 1: AGT mistargeting highlights the fundamental differences between the peroxisomal and mitochondrial protein import pathways Biochim Biophys Acta 1763, 1776–1784

3 Salido EC, Li XM, Lu Y, Wang X, Santana A, Roy-Chowdhury N, Torres A, Shapiro LJ &

Roy-Chowdhury J (2006) Alanine-glyoxylate amino-transferase-deficient mice, a model for primary hyperoxaluria that responds to adenoviral gene transfer Proc Natl Acad Sci USA 103, 18249–18254

4 Singh H, Beckman K & Poulos A (1994) Peroxisomal beta-oxidation of branched chain fatty acids in rat liver Evidence that carnitine palmitoyltransferase I prevents transport of branched chain fatty acids into mitochon-dria J Biol Chem 269, 9514–9520

5 Li XM, Salido EC & Shapiro LJ (1999) The mouse alanine:glyoxylate aminotransferase gene (Agxt1): cloning, expression, and mapping to chromosome 1 Somat Cell Mol Genet 25, 67–77

6 Pilkis SJ & Granner DK (1992) Molecular physiology

of the regulation of hepatic gluconeogenesis and glycolysis Annu Rev Physiol 54, 885–909

7 Noguchi T, Okuno E, Takada Y, Minatogawa Y, Okai K & Kido R (1978) Characteristics of hepatic alanine-glyoxylate aminotransferase in different mammalian species Biochem J 169, 113–122

8 Baker PR, Cramer SD, Kennedy M, Assimos DG & Holmes RP (2004) Glycolate and glyoxylate metabolism

in HepG2 cells Am J Physiol Cell Physiol 287, C1359– C1365

9 Xue HH, Fujie M, Sakaguchi T, Oda T, Ogawa H, Kneer NM, Lardy HA & Ichiyama A (1999) Flux of the L-serine metabolism in rat liver The predominant contribution of serine dehydratase J Biol Chem 274, 16020–16027

10 Xue HH, Sakaguchi T, Fujie M, Ogawa H & Ichiyama

A (1999) Flux of the L-serine metabolism in rabbit, human, and dog livers Substantial contributions of both mitochondrial and peroxisomal

274, 16028–16033

11 Neuhoff V, Arold N, Taube D & Ehrhardt W (1988) Improved staining of proteins in polyacrylamide gels including isoelectric focusing gels with clear background

at nanogram sensitivity using Coomassie Brilliant Blue G-250 and R-250 Electrophoresis 9, 255–262

12 Mortz E, Krogh TN, Vorum H & Gorg A (2001) Improved silver staining protocols for high sensitivity protein identification using matrix-assisted laser

Proteomics 1, 1359–1363

13 Sambrook J & Russell DW (2001) Molecular Cloning:

A Laboratory Manual Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY