Cloning, Sequencing and Tissue Distribution of Rat Flavin-Containing Monooxygenase 4: Two Different Forms Are Produced by Tissue-Specific Alternative Splicing

  Introduction

Top
Abstract
Introduction
Materials and Methods
Results
Discussion
References

Microsomal flavin-containing monooxygenases (FMOs) [dimethylaniline monooxygenase (N-oxide forming); EC 1.14.13.8] catalyze the FAD-, NADPH- and O₂-dependent oxidation of a large number of structurally diverse compounds, including drugs, pesticides, and industrial chemicals containing a soft nucleophile (Ziegler, 1988). The FMOs convert many xenobiotics into more polar substances as a prelude to excretion. In certain cases, the FMOs can also catalyze the formation of reactive metabolites capable of binding to cellular macromolecules. Several endogenous compounds such as trimethylamine, methionine, or cysteamine (Sausen and Elfarra, 1990; Duescher et al., 1994; Lang et al., 1998) have been identified as substrates for FMOs, but the physiological role of some of these enzymes has not been determined.

To date, a maximum of five distinct FMO isoforms have been identified. Based on the cDNA sequence, they were classified into five subfamilies (FMO1 to 5) (Lawton et al., 1994). Orthologous sequences share at least 80% of amino acid identity, whereas homologous FMOs are 52 to 57% identical. The FMO isoforms differ in their tissue distribution or substrate specificity. The expression of recombinant proteins has greatly aided the characterization of FMOs isoforms. So, the cDNAs of FMO1, FMO2, FMO3, and FMO5 isoforms of various animal species have been expressed successfully in several expression systems (Lawton et al., 1991, 1993; Itoh et al., 1997) and high levels of expression have been observed in Escherichia coli (Atta-Asafo-Adjei et al., 1993; Lawton and Philpot, 1993; Lomri et al., 1993; Overby et al., 1995; Falls et al., 1997).

In contrast to these successes, the expressions of rabbit or human FMO4 in E. coli, yeast, or COS-1 cells have always failed for unknown reasons (Burnett et al., 1994; Phillips et al., 1995; Itagaki et al., 1996). The comparison of the human (Dolphin et al., 1992) and rabbit (Burnett et al., 1994) FMO4 transcripts with those encoding all known FMO isoforms, revealed that the transcripts of the coding regions of FMO4 are 60 to 75 nucleotides longer than the transcripts of the coding regions of other isoforms. This has been suspected to be caused by a shift in the stop codon to the 3'-end of the consensus position. The supplementary C-terminal extremity could be associated with the incapacity to express the FMO4s in the in vitro expression systems, because the expression of a truncated FMO4 that does not contain the supplementary C-terminal extremity has been obtained in E. coli without real difficulty.

The fact that no purified FMO4 enzyme has been obtained up to now explains that no FMO4 antibodies are available, the tissue expression of FMO4 protein could not be recognized, and the catalytic activity of this protein could not be characterized. There are also no endogenous or exogenous substrates known for FMO4 at the present time. Nevertheless, the distribution of mRNA suggests that FMO4 plays some functional role in the kidney (Burnett et al., 1994) and brain (Blake et al., 1996).

In this article, we report for the first time the complete sequence of the cDNA encoding for rat FMO4, the existence of a tissue specific alternative splicing, and the tissue-specific expression of two FMO4 variants.

  Materials and Methods

Top
Abstract
Introduction
Materials and Methods
Results
Discussion
References

Animals. Male or female Oncins France strain A/Sprague-Dawley rats (200 g) were obtained from a commercial breeder (IFFA-CREDO, L'Arbresle, France). Food and water were made available to rats ad libitum until sacrifice. Rats were killed by decapitation. Various organs (kidney, brain, liver, lung, intestine, adrenal gland, ovary, testis, uterus) were removed quickly, frozen in liquid nitrogen and stored at 80°C until analysis.

Standard Procedures. Standard molecular biology techniques were carried out essentially as described by Sambrook et al. (1989). Poly(A)⁺ RNA was purified from a fresh adult rat liver (100 mg) with streptavidin-coated paramagnetic particles (PolyATract mRNA isolation system; Promega, Charbonnières, France). Total RNA were isolated from rat liver, kidney, lung, brain, adipose tissue, and muscle by a single extraction with an acid guanidinium thiocyanate/phenol/chloroform solution according to Chomczynski and Sacchi (1987). Oligonucleotides were manufactured by OligoExpress (Paris, France). Sequencing was performed by Genome Express (Grenoble, France).

Amplification of an Internal Fragment of the cDNA Encoding the Rat FMO4. The first strand cDNA template was synthesized from rat kidney mRNA (50 ng) in the presence of oligo(dT)₁₅ (100 pmol), 200 units of Moloney murine leukemia virus reverse-transcriptase RNase H (MMLV-RT; Promega) in 20 µl of standard reverse transcription buffer (50 mM Tris-HCl, pH 8.3, 3 mM MgCl₂, 75 mM KCl, 10 mM dithiothreitol, and 200 µM concentrations of each deoxynucleotide triphosphate) at 37°C for 1.5 h and then at 65°C for 10 min.

Amplification of FMO4 cDNA Ends. To obtain the cDNA ends, the SMART RACE cDNA amplification kit (BD Clontech, Palo Alto, CA) was used according to the manufacturer's recommendations. The specific primers (FMO4-AS2, FMO4-AS3, FMO4-S2) used were designed from sequences of clones F4-1 to F4-3.

Amplification of the Full-Length cDNA Encoding Rat FMO4. A (dT)₁₅ primed first-strand cDNA template was synthesized from rat kidney or brain mRNA (50 ng) in the presence of oligo(dT)₁₅, in the same conditions as described above. Sense FMO4-S4 (5'-GACCGAAACTTCTTCCTGTTG-3') and antisense FMO4-AS4 oligonucleotides (5'-AACCTGC CCTATGGAGACTG-3') designed from the 5'- and 3'-ends nonencoding sequences (from base 222 to base 201 and from base 1778 to 1759), respectively, were used to perform the amplification, by PCR, of the complete coding region. After an initial denaturation step at 94°C for 2 min, 1.25 units of Pfu DNA polymerase (Promega) were added and the reaction was incubated for 35 cycles at 94°C for 30 s, 58°C for 30 s, and 72°C for 4 min and for an additional 10 min at 72°C. The amplified products were further submitted to PCR performed using sense FMO4-S5 (5'- GACAACTTCAGACCTTCACCC-3') and antisense FMO4-AS5 (5'-AACTTCCTCCATTGCCAAAC-3') primers in the same conditions (from base 125 to 105 and from base 1710 to 1691, respectively). The resulting products (1600 and 1800 bp) were gel-purified, cloned into pSTBlue-1, and sequenced (clones F4L-4 to -9 and clones F4C-10 to -12).

Amplification of a Region of the FMO4 Gene. Genomic DNA was isolated from Sprague-Dawley rat kidney by use of a commercial DNA isolation kit (DNeasy tissue kit; QIAGEN, Valencia, CA). A region of the FMO4 gene was amplified by PCR using the sense FMO4-S6 (5'-AGAGGAGTAGTGACTTCGGCG-3') and the antisense FMO4-AS6 (5'-TGTGACAACTTCCCATTGGC-3') primers (from base 417 to 437 and from base 709 to 690, respectively). Reaction mixtures containing approximately 200 ng of genomic DNA as template were incubated in the presence of Advantage polymerase mix (BD Clontech) for 25 cycles at 94°C for 45 s, 58°C for 45 s, and 72°C for 4 min, followed by an additional 10 min at 72°C. The PCR products were gel purified and sequenced directly.

Northern Blotting. Twenty micrograms of total RNA isolated from each organ were separated by electrophoresis on agarose gel (1.2%) containing formaldehyde (2 M) and then transferred to nylon membrane (Hybond N; Amersham Biosciences AB, Umeå, Sweden). The blot was washed in 5× SSC for 15 min at room temperature and baked at 80°C for 2 h. The membrane was prehybridized for 4 h at 42°C and then hybridized with a rat FMO4 coding region probe (EcoRI/EcoRI, bases 280 to 981 of long FMO4 cDNA) at 42°C for 24 h. The probe was labeled by a random primer method (Prime-a-Gene labeling system; Promega). The hybridized membrane was washed for two cycles of 10 min in 2× SSC (0.1% SDS), then two cycles of 15 min in 1× SSC (0.1% SDS) at room temperature, and at 55°C for two cycles of 15 min in 0.1× SSC (0.1% SDS). The membrane was subjected to autoradiography for 48 h at 80°C. The membrane was stripped with a 0.01× SSC (0.5% SDS) at 100°C. The removal of the probe was verified by autoradiography at 80°C. The blot was subsequently hybridized with a actin cDNA probe.

Semiquantitative RT-PCR. Total RNA were extracted from various organs (liver, kidney, lung, uterus, intestine, whole brain homogenate, brain stem, cerebellum, cortex, hypothalamus, hippocampus, adrenal glands, testis, and ovary) by a double extraction, beginning with a manual extraction with an acid guanidinium thiocyanate/phenol/chloroform solution followed by an extraction on column (total RNA isolation system, Promega). Concentrations of total RNAs were evaluated spectrophotometrically from absorbance at 260 nm. The 260/280 ratios were between 1.8 and 2.0. The reverse transcription of 0.5 µg of total RNA was performed in the presence of oligo(dT)₁₅ (500 pmol) in 20 µl of standard reverse transcription buffer, as described above. Two microliters of resulting cDNA were amplified by PCR using specific primers FMO4-S4 and FMO4-AS7 (5'-GGCATCTGGAATCCTGTACT-3', from base 543 to 524) in 50 µl of standard PCR buffer. The amplification was performed at 94°C for 90 s, 24 to 35 cycles of 94°C for 30 s, 61°C for 45 s, and 72°C for 60 s. The semiquantification of FMO4 transcripts was estimated by the number of amplification cycles necessary to obtain a similar signal density.

Production of Specific Antibodies Raised against Rat FMO4 Protein. Two oligopeptides, the one corresponding to amino acid residues 126 to 140 of deduced long FMO4 protein (P1), the other one corresponding to amino acid residues 409 to 423 of deduced long FMO4 protein (P2) were selected. These peptides have reasonable hydrophilic characters and specific sequences to rat FMO4 compared with known FMOs. The sequences of these peptides were EVVTETEGKQDRAVFY and YKKEELIKRGVIKDIS for peptides 1 and 2, respectively; the terminal tyrosines were not part of the rat FMO4 sequence but were added to facilitate coupling to the carrier protein. The oligopeptides were synthesized by Neosystem (Strasbourg, France). The composition of peptides was verified by mass spectrometry and amino acid analysis. Purity of the peptides was verified by reversed-phase high-performance liquid chromatography using a C-18 column (Neosystem). The oligopeptides were coupled to the carrier protein of keyhole limpet hemocyanin (KLH) by Neosystem. The peptides-KLH were used as an immunogen to immunize rabbits. The initial injection consisted of 400 µg of oligopeptide 1 or 2/KLH mixed with an equal volume of Freund's complete adjuvant (Sigma, St. Louis, MO); the immunogen was injected intradermally dispersed among 20 to 30 sites on the back of the rabbit. Intradermal booster injections of 200 µg of oligopeptide 1 or 2/KLH mixed with an equal volume of Freund's incomplete adjuvant (Sigma) were administered at 14, 28, and 56 days after the initial injection. Animals were bled 10 days after the last booster injection. Sera were tested for their reactivity to the recombinant rat long FMO4 by Western blotting.

Heterologous Expression in E coli The long and short forms of rat FMO4 were expressed in the glutathione S-transferase (Schistosoma japonicum) gene fusion vector, pGEX-6P3 (Amersham Biosciences AB). The coding sequence corresponding to the long and short FMO4 were amplified, by PCR, from clones F4L-4 and F4C-12, using Pfu DNA polymerase and specific primers FMO4S-BamHI (5'-ATAGGATCCATGGCCAAAGAAAGTGGCAG-3') and FMO4AS-SalI (5'-ACAGTCGACTCATGAGTTCTGGGGGATGTAC-3'), including BamHI and SalI restriction sites in the extremities, respectively. After an initial denaturation step at 94°C for 2 min, 1.25 units of Pfu DNA polymerase were added and the reactions were incubated for 5 cycles at 94°C for 30 s, 50°C for 30 s, and 72°C for 4 min, for 25 cycles at 94°C for 30 s, 65°C for 30 s, and 72°C for 4 min and then for an additional 10 min at 72°C. The gel-purified PCR products (1680 or 1490 bp) were digested by 10 units of BamHI and SalI restriction enzymes (Roche Diagnostics, Meylan, France) and ligated into the expression vector pGEX-6P3 linearized by the same enzymes. The clones, named pGEX-F4L and pGEX-F4C, were obtained and verified by sequencing.

In Vitro Transcription/Translation. The in vitro transcription/translation of long and short FMO4s was performed by use of a commercial kit (TNT T7 Quick for PCR DNA; Promega). The PCR fragments were produced from clones pGEX-F4L and pGEX-F4C in the presence of primers T7-FMO4S (5'-TTATAATACGACTCACTATAGGGAGCCACCATGGCCAAGAAAGTGGCAGTG-3') and FMO4AS-SalI. The PCR products (1680 bp for long FMO4 and 1491 bp for short FMO4) were gel-purified, then added (150 ng of PCR product) in the translation reactions, according to the manufacturer's recommendations. After an incubation for 1 h at 30°C, the reaction products were analyzed by Western blot in the presence of antibodies raised against P1 or P2.

Analysis of Proteins Expression. Microsomes were prepared from rat liver, lung, kidney, brain, intestine, and uterus samples by differential centrifugation as described previously (Moroni et al., 1995). Proteins concentrations were determined by the method of Bradford (1976) using serum bovine albumin as a standard. Microsomal proteins (30 µg) were loaded, and immunoblottings were performed as described above. Membranes were blocked and then incubated with antibodies anti-P1 (1:5000) or anti-P2 (1:2500). After washing, membranes were incubated with horseradish peroxidase-conjugated anti-rabbit IgG (Amersham Biosciences AB), as the secondary antibody. FMO4 was visualized by enhanced chemiluminescence on Hyperfilm (Amersham Biosciences AB).

  Results

Top
Abstract
Introduction
Materials and Methods
Results
Discussion
References

PCR-Based Cloning of cDNA. RT-PCR performed with rat kidney mRNA with sense primer FMO4-S1 and antisense primer FMO4-AS1, as described under Materials and Methods, generated a 670-bp fragment on an ethidium bromide-stained agarose gel (results not shown). This 670-bp fragment was sequenced and compared with human (Dolphin et al., 1992) and rabbit (Burnett et al., 1994) FMO4 cDNAs. The results showed 82 and 83% homology with human and rabbit cDNAs, respectively.

Sequencing of Two Different cDNAs Encoding the Long and Short FMO4. RT-PCR performed with rat kidney mRNA, with specific primers FMO4-S4 and FMO4-AS4 designed from the 5'- and 3'-ends nonencoding sequences of rat FMO4, generated an 1800-bp fragment. The 1800-bp product thus obtained was cloned and sequenced from three different Sprague-Dawley rat kidney mRNAs. The three clones showed very limited differences (from 1 to 3 bases). The consensus nucleotide sequence  for the rat FMO4 cDNA obtained by alignment of these three sequences contains 322 bases at the 5'-flanking region followed by an open reading frame of 1680 bases, terminated with a TGA stop codon, and 380 bases at the 3'-flanking region . The obtained nucleotide sequences showed 81 and 80% homology with human (Dolphin et al., 1992) and rabbit (Burnett et al., 1994) FMO4 cDNAs.

Long and Short FMO4 Transcripts Are Produced by Alternative Splicing. PCR amplification in the presence of FMO4-S6 and FMO4-AS6 primers was performed with rat kidney and brain cDNA. A single 293-bp fragment was amplifed from rat kidney cDNA, whereas two PCR products (293 and 104 bp) were amplified from rat brain cDNA. The same amplification using the rat genomic DNA generated a single PCR fragment containing about 2200 bp . After gel purification, this fragment was directly sequenced. Comparison of the sequences obtained from amplification of genomic DNA with the transcript sequence obtained from the long FMO4 cDNA revealed the presence of two intronic sequences framing the 189-bp region specific of the long transcript in the FMO4 gene. The first intron, found between bases 132 and 133 of the long transcript cDNA, was composed of 1360 bp. It contained all the characteristic sites of intron (5'- GT... ... ... . CTTCTAAC(-21)... . TATAG -3'). The other intron, found between bases 321 and 322 of the long transcript cDNA, contained 541 bp. The 189-bp region specific to the long transcript corresponded to a unique exon that was spliced out only in the short transcript.

Tissue Distribution of Long and Short FMO4 mRNAs in the Rat. Tissue distribution of long and short mRNA in various rat tissues was examined by Northern blot  and RT-PCR. Expression of mRNAs encoding FMO4 obtained in the male and female rat liver (lanes 1 and 2), kidney (lanes 3 and 4), lung (lanes 5 and 6), brain (lanes 7 and 8), muscle (lanes 9 and 10), adipose tissue (lanes 11 and 12), adrenal gland (lanes 13 and 14), testis (lane 15), ovary (lane 16), and intestine (lanes 17 and 18) is shown in . The probe for FMO4 specifically hybridized a 2.8-kilobase transcript in the intestine and kidney samples . FMO4 mRNA levels were higher in the kidney compared with the intestine. No FMO4 mRNA signal was detected in other tissues, although the mRNAs integrity was confirmed by hybridization with an actin cDNA probe. Because the Northern blot sensitivity was lower than the RT-PCR sensitivity, the presence of the FMO4 mRNA was studied using a semiquantitative RT-PCR, in the presence of FMO4-S4 and FMO4-AS7, framing the 189-bp deletion. The number of PCR cycles necessary to obtain a similar band signal intensity on an ethidium bromide-stained agarose gel was used as an index of the mRNA abundance. The kidney was again found to contain a high level of FMO4 mRNA, needing only 24 PCR cycles to demonstrate its constitutive expression. The intestine and liver contained a more limited amount of FMO4 mRNA (28 PCR cycles were necessary) followed by uterus and adrenal gland (30 PCR cycles), then the whole-brain homogenate (32 PCR cycles), which was followed by testis, ovary, and lung (34 PCR cycles). In the rat brain, the constitutive expression of short form FMO4 mRNA was higher in the brain stem and hypothalamus (30 PCR cycles) than that observed in the hippocampus, cortex, or cerebellum (34 PCR cycles). Among the various organs tested, the PCR performed with the cDNA obtained either from whole brain homogenate or the homogenates of individual specific areas of the brain generated both transcripts (765 and 576 bp). The long-form FMO4 mRNA transcript was always more abundant than the short-form FMO4 mRNA transcript 

Reactivity of Antipeptide Antibodies. cDNAs encoding long and short FMO4 were cloned into E. coli expression vectors (pGEX-6P3). The IPTG-dependent expression of pGEX-FMO4 in the E. coli BL21 was controlled by the measurement of the glutathione S-transferase activity in the sonicates. Low GST activity was detected in the sonicates containing long FMO4 or short FMO4. The presence of recombinant proteins was analyzed by Western blot using the antibodies raised against P1 or P2. Only the analysis of the sonicate containing the GST-long FMO4 fusion protein demonstrated the presence of a band, recognized by both antibodies, at the expected molecular mass (~90 kDa). No band migrating at about 85 kDa was detected in the sonicate containing short FMO4. In contrast, numerous bands of lower molecular mass were recognized by both antibodies in the sonicates containing either the long or the short FMO4 (result not shown). These lower bands seem to be the result of a strong proteolysis during the IPTG-dependent expression. The short FMO4 not having been obtained, only the long FMO4 was purified. The proteins were solubilized from inclusion bodies of E. coli using N-lauroyl-sarcosine (1.5%), because the induced GST-long FMO4 fusion protein was insoluble in Triton X-100 (2%). This step may have compromised the ability to recover functional long FMO4. Solubilized fusion proteins were bound on a glutathione Sepharose 4B. The long FMO4 protein was eluted by cleavage with Prescission Protease, which keeps GST remaining on the column. The analysis of the purified long FMO4 by Western blot using antibodies raised against P1 or P2 revealed the presence of a band migrating at ~64 kDa  along with several other bands with lower molecular mass, seeming to be the result of the proteolysis as described above.

Expression of Long FMO4 in Kidney and Intestine and Short FMO4 in the Brain. Microsomes were prepared from kidney, intestine, uterus, liver, lung and brain of rats, and the constitutive expression of either forms of FMO4 was analyzed by Western blotting using antibodies raised against P1  or P2  A 64-kDa protein corresponding to the long FMO4 was detected in the rat kidney  and intestine (result not shown). This 64-kDa protein was not detected in the rat brainor uterus (result not shown). On the other hand, a faint 57-kDa band corresponding to the short FMO4 was recognized in the rat brain . Neither the 64-kDa protein nor the 57-kDa protein were detected in the liver or in the lung (results not shown).

  Discussion

Top
Abstract
Introduction
Mterials and Methods
Results
Discussion
References

In this study, a cDNA clone encoding a protein of 560 amino acids was obtained from rat kidney by RT-PCR and 5'/3' terminal extension, using oligonucleotides designed from the sequences of FMO4s previously published. The deduced amino acid sequence revealed high homology with the FMO4s of humans (80%) (Dolphin et al., 1992) and rabbits (79%) (Burnett et al., 1994), whereas the homology was lower compared with the sequences of rat FMO1 (50%) (Itoh et al., 1993) or rat FMO3 (51%) (Lattard et al., 2001). The designation "rat FMO4" for this cDNA is in agreement with the actual nomenclature based on the comparison of the amino acid sequence of the mammalian FMOs (Lawton et al., 1994).

A 1800-bp cDNA containing an ORF of 1680 bp was obtained from rat kidney mRNA. The sequence surrounding the proposed initiation codon GCACCATGG is favorable for translational initiation of vertebrate mRNAs (Kozak, 1987). A polyadenylation signal (AATAAA) was found 22 bases upstream of a polyadenylated tail. From rat brain mRNA, two different transcripts (1800 and 1650 bp, respectively) encoding both long and short forms of FMO4 were amplified, cloned, and sequenced. The long transcript was identical to that previously sequenced in the rat kidney. The short transcript, encoding a protein of 497 amino acid residues, is the result of a 189-bp deletion, corresponding to the nucleotides from 132 to 321 of rat kidney FMO4 cDNAs. The constitutive presence of these two transcripts in the rat brain was confirmed by a PCR amplification of rat brain total RNA using specific primers framing the 189-bp deletion. Furthermore, the expression of short-form transcript seems to be specific in the rat brain. The constitutive expression of both forms of FMO4 transcripts is found in all cerebral regions tested.

The amplification of the corresponding region of genomic DNA revealed the presence of a single FMO4 gene encoding the long form FMO4. The short FMO4 seems to be caused by an alternative splicing specific of the brain. The deleted 189-bp region corresponds exactly to the exon, which corresponds to exon 4 present in the human FMO4 gene (GenBank accession number ). Exon 4 of the rat FMO4 has some intronic characteristics. The sequence AG observed at the 3' extremity of the deleted region corresponds to a potential acceptor site (Mount, 1982; Shapiro and Senapathy, 1987; Stephens and Schneider, 1992) for alternative splicing, and the sector rich in pyrimidine bases at 18 bases upstream could be considered as a potential branch site. On the other hand, the 5' extremity of the exon 4 contains GA instead of GT, which is characteristic of a 5' splice donor (Mount, 1982; Shapiro and Senapathy, 1987; Stephens and Schneider, 1992). However recent studies of the thioredoxin reductase 1 genes (Osborne and Tonissen, 2001) related the presence of the 5' splice donor GC instead of GT, which is probably involved in the alternative splicing of the 5' region. The occurrence of this modified splice signal could represent a potential regulation toward formation of the short form FMO4. The sequence of the exon 4 of the human FMO4 gene is similar to the rat FMO4 exon 4. The existence of such an alternative splicing in human tissues should be explored.

Similar to the forms of FMO described previously, the long sequence exhibits the presence of characteristic FAD- and NADPH-binding sites, beginning at residues 9 and 191, respectively. As for the previously described forms of FMO4 (Dolphin et al., 1992; Burnett et al., 1994), the deduced protein is approximately 28 residues longer than other FMO isoforms and has a calculated molecular mass of ~64 kDa. These additional residues present in FMO4 are contained in a single block located at the COOH terminus and may have resulted from the loss of the stop codon commonly present in all other FMO isoforms. The translation continues to the following stop codon, localized 84 nucleotides farther on. Because of this extension, rat FMO4 terminates with an additional hydrophilic peptide consisting of 19 extra amino acids not present in other FMO isoforms (Ozols, 1991). The 189-bp deletion does not modify the reading frame of the short sequence and corresponds to the loss of 63 amino acid residues (from residue 44 to 117) of the long sequence. Despite this deletion, putative FAD- (amino acids from 9 to 14) and NADPH- (amino acids from 128 to 133) binding domains were preserved. Nevertheless, the hydrophilic region localized usually between these two sites is not found in the short FMO4 form, perhaps expressed constitutively in the brain. Such exclusion of exon 4 in the short FMO4 form may constitute a significant structural and functional change.

The corresponding mRNA was found, by Northern blot, only in the rat kidney and intestine only. The constitutive expression of FMO4 mRNA in the kidneys of rabbit, guinea pig, or rat kidney had been already suggested by Northern blot studies using a rabbit cDNA probe (Burnett et al., 1994). In the rat kidney FMO1 (Itoh et al., 1993) and FMO3 (Lattard et al., 2001) are also found. The presence of FMO4 mRNA was detected in the liver of the guinea pig but not in the rat liver (Burnett et al., 1994). Our results obtained both by Northern blot and RT-PCR confirm that the expression of FMO4 mRNA in the rat liver is, for the best, extremely limited. FMO4 mRNAs (both short and long transcripts) were detected in the rat brain by RT-PCR only. This observation is coherent with the presence of FMO4 transcripts, detected by RT-PCR in the rabbit brain (Blake et al., 1996). All the tested cerebral regions contained these mRNA but the expression was higher in the hypothalamus and brain stem. The alternative splicing of the mRNA encoding for mRNA seems to be specific to the brain.

To date, the tissue expression of the corresponding FMO4 protein has not been reported. The production of rabbit or human recombinant proteins has always failed (Burnett et al., 1994; Phillips et al., 1995; Itagaki et al., 1996), with the exception of a truncated human FMO4 (Itagaki et al., 1996). Because of these difficulties, probably linked to a strong proteolysis, the production of even a limited amount of FMO4 necessary to immunize rabbit has never been achieved. Consequently, no production of FMO4 antibodies has been reported. Our attempts to produce rat FMO4 were also partially unsuccessful. Only the long FMO4 was produced, in very small quantities and strongly degraded. Consequently, we could produce antibodies against only two rat FMO4 peptides. The limited amount of available recombinant long FMO4 was, however, sufficient to verify the recognition of the rat FMO4 by employing the two antibodies raised against the peptide fragments of long-form FMO4. The reactivity of these antibodies against long and short protein obtained by in vitro transcription/translation was also verified. The absence of cross-reactivity toward other FMO isoforms is confirmed by the absence of signal in Western blot between 40 and 70 kDa in rat liver and lung. Indeed FMO3, FMO1, and probably FMO5 are highly expressed in the liver (Lattard et al., 2001), and FMO2 is found in the lung (Lattard et al., 2002). The use of these two antibodies demonstrated the expression of a long FMO4 protein in the rat kidney and intestine but not in the rat brain. On the other hand, the expression of a short FMO4 protein in the rat brain was demonstrated and confirmed by both antipeptide antibodies.

Because a large number of psychotropic and neurotoxic agents are metabolized by FMO, the presence of FMO in the brain has been suspected for a long time. Metabolism of FMOs substrates (thiobenzamide, methimazole, dimethylaniline, or benzydamine) has already been reported in the rat brain (Duffel and Gillepsie, 1984; Bhamre and Ravindranath, 1991; Bhamre et al., 1993; Kawaji et al., 1994). But the expressed isoforms have not yet been identified. The expression of FMO1 (Itoh et al., 1993) or/and FMO2 (Bhamre et al., 1993) were suggested, but not confirmed. In the other hand, the presence of FMO4 transcript in the rabbit brain was reported by Blake et al. (1996). The expression, described for the first time in this study, of a short FMO4 protein in the rat brain produced by tissue specific alternative splicing, and the expression of a long FMO4 protein in the kidney, could be of great interest. Nevertheless, further investigations concerning the catalytic properties of these proteins are necessary.