Intergenic mRNA molecules resulting from trans-splicing.

Accumulated recent evidence is indicating that alternative splicing represents a generalized process that increases the complexity of human gene expression. Here we show that mRNA production may not necessarily be limited to single genes, as human liver also has the potential to produce a variety of hybrid cytochrome P450 3A mRNA molecules. The four known cytochrome P450 3A genes in humans, CYP3A4, CYP3A5, CYP3A7, and CYP3A43, share a high degree of similarity, consist of 13 exons with conserved exon-intron boundaries, and form a cluster on chromosome 7. The chimeric CYP3A mRNA molecules described herein are characterized by CYP3A43 exon 1 joined at canonical splice sites to distinct sets of CYP3A4 or CYP3A5 exons. Because the CYP3A43 gene is in a head-to-head orientation with the CYP3A4 and CYP3A5 genes, bypassing transcriptional termination can not account for the formation of hybrid CYP3A mRNAs. Thus, the mechanism generating these molecules has to be an RNA processing event that joins exons of independent pre-mRNA molecules, i.e. trans-splicing. Using quantitative real-time polymerase chain reaction, the ratio of one CYP3A43/3A4 intergenic combination was estimated to be approximately 0.15% that of the CYP3A43 mRNAs. Moreover, trans-splicing has been found not to interfere with polyadenylation. Heterologous expression of the chimeric species composed of CYP3A43 exon 1 joined to exons 2-13 of CYP3A4 revealed catalytic activity toward testosterone.

The ongoing genome projects are allowing the determination of the coding potential of several eukaryotic species at a rapidly increasing pace. However, as most eukaryotic genes are in pieces, consisting of discrete exonic sequences flanked by noncoding intronic segments, defining the expressed fraction of the genetic material is obviously of highest priority. The puzzling problem of identifying relatively short exons in a vast excess of intronic sequences is further complicated by recent results indicating that, during gene expression, an additional level of complexity also exists. Specifically, in several higher eukaryotes, processed RNA does not always represent a simple linear combination of exonic sequences. Thus, in addition to canonical transcripts, gene expression may also result in scrambled RNA molecules in which exons juxtapose in an order that is different to that in the gene (1)(2)(3)(4)(5)(6)(7)(8)(9)(10). Moreover, the continuously increasing number of reports providing evidence for intergenic RNA molecules suggests that RNA splicing may also combine exons originating from more than one gene (Ref. 10 and references therein; see also . Because a majority of these hybrid transcripts is known to encompass exons from neighboring genes that have the same orientation, a plausible mechanism for intergenic mRNA formation would involve bypassing transcriptional termination and formation of a bicistronic transcript followed by alternative splicing. However, a number of in vivo and in vitro experiments indicated that the mammalian splicing apparatus also has the capability to join exons originating from distinct pre-mRNA molecules, a process termed trans-splicing (15)(16)(17)(18)(19). The recent findings of exon repetitions in some rat and human mRNAs provided the most convincing evidence that natural trans-splicing within transcripts of single genes may also occur in mammalian cells (20 -23).
Cytochrome P450 3As (CYP3A) 1 are heme-containing monooxygenases that catalyze stereospecific hydroxylation of a wide range of substrates, including endogenous steroids and more importantly various xenobiotics, such as drugs and environmental chemicals (24). There are four known CYP3A genes in humans, CYP3A4, CYP3A5, CYP3A7, and the recently discovered CYP3A43 (25)(26)(27). The proteins are very similar to each other bearing 71-88% amino acid identities. All four genes consist of 13 exons and form a cluster on chromosome 7q21-22.1. Interestingly, comparison of GenBank entries encompassing various CYP3A genomic sequences revealed a rather unique arrangement of the CYP3A locus, specifically that the CYP3A43 gene is in a head-to-head orientation with the other three, CYP3A4, CYP3A7, and CYP3A5, genes (Ref. 25; Fig. 1).
Recently we have shown that expression of members of another group of cytochrome P450s, the CYP2C, results, in addition to canonical mRNAs, in a variety of chimeric RNA molecules in human liver and epidermis (10,28). Here we provide evidence that chimeric RNA production is not limited to the CYP2C genes on chromosome 10q24, but is also typical of the CYP3A genes. Specifically, in liver several hybrid mRNAs are produced that encompass the first exon of CYP3A43 joined to various sets of either CYP3A4 or CYP3A5 exons. In addition, because the CYP3A43 gene has an opposite orientation to the CYP3A4 and CYP3A5 genes, the mechanism of chimeric mRNA formation presumably involves splicing events between CYP3A43 and either CYP3A4 or CYP3A5 pre-mRNA molecules. Thus, these findings suggest that natural trans-splicing between distinct genes may occur in higher eukaryotes. Moreover, this intricate pattern of expressed mRNAs implies that novel alternative splicing pathways may significantly increase the complexity of human gene expression.

MATERIALS AND METHODS
Reverse Transcription-Polymerase Chain Reaction-Human total RNA was purchased from Invitrogen (adult human liver) or purified from HepG2 cells. HepG2 cells (ATCC) were grown at 37°C, in an atmosphere of humidified air containing 5% CO 2 , in RPMI 1640 with L-glutamine medium (Invitrogen) supplemented with 10% fetal calf serum (Invitrogen), 100 units/ml penicillin, and 100 g/ml streptomycin (Invitrogen). Subculturing was performed at a subcultivation ratio of 1:6. Cells from confluent cultures were harvested by addition of trypsin, and subjected to RNA preparation using the SV Total RNA Isolation System (Promega).
cDNA synthesis was carried out with Moloney murine leukemia virus reverse transcriptase (Promega) in a 50-l reaction mixture using oligo(dT) primers (25 ng/ml, Promega) with 10 g of total RNA template. An aliquot of 1-5 l of the reverse transcription reaction was directly subjected to the first amplification. Both the initial and nested polymerase chain reactions (PCRs) were performed in a 50-l reaction mixture containing 20 pmol of each of the forward and reverse primers (Cybergene AB), 2 mM MgCl 2 , 200 M dNTPs, and 1.25 units of Taq DNA polymerase (Promega). The oligonucleotides used in the PCRs are listed in Table I. The reactions were carried out for 30 cycles, with 10 s at 92°C, 30 s at 52°C, and 2 min at 72°C. For the nested reaction, 1 l of the first PCR was directly used. Alternatively, amplifications were done by using the Expand Long Template PCR System (Roche Molecular Biochemicals) for 30 cycles with 10 s at 92°C, 30 s at 52°C, and 2 min ϩ5 s/cycle at 68°C.
Analysis of PCR Products-The PCR products were cloned in Escherichia coli XL1-Blue cells using the pGEM-T Vector System (Promega). The cloned PCR products were sequenced using the BigDye Terminator Cycle Sequencing Ready Reaction Kit (Applied Biosystems). Electrophoresis of the sequenced samples was performed by Cybergene AB or KISeq (Karolinska Institute). The analysis of the DNA sequences was done with the BLAST similarity search program (29).
RNase Protection Assay-RNA was isolated from a human liver sample (a generous gift of Dr. Nikos Papadogiannakis, Huddinge University Hospital) by the guanidinium thiocyanate method (30). A DNA construct containing a segment of the (1) 3A43 -(2-3-4-5-6-7-8-9-10-11-12-13) 3A4 cDNA spanning 43 base pairs (bp) upstream and 25 bp downstream of the nontypical splice junction was generated by PCR amplification (10 cycles) of that cDNA with Advantage HF polymerase (CLONTECH) and primers 5Ј-GCG CGG CCG CAA ACA TGG GTT CTT GTG GC and 5Ј-GCG TCG ACA AGT CCA TGT GAA TGG GTT CCA TAT (Cybergene AB). The presence of NotI and SalI sites in the primers allowed the directional cloning of the PCR product into the pGEM-5Z (Promega) vector, and the sequence of the construct was confirmed by direct dye-deoxy sequencing. The construct was linearized with NotI (Promega) and transcribed (MAXIscript, Ambion) in the presence of 32 P-labeled UTP (Amersham Biosciences, Inc.) with either SP6 polymerase to generate the riboprobe or with T7 polymerase to produce a molecular weight marker for the analysis of the protected fragments. 50 and 150 g of total RNA from human liver were incubated with equal amounts (2 ϫ 10 5 cpm) of the riboprobe and subjected to RNase protection following the protocol of the RPA III kit (Ambion). Electrophoretic analysis of the riboprobe and the protected fragments was performed on a denaturing 10% acrylamide, 8 M urea gel and visualized after exposure to Fuji Super RX film.
Southern Hybridization-Aliquots (10 g) of human genomic DNA (Promega) were digested individually with BglII, EcoRI, NdeI, and SpeI endonucleases (Promega). Following electrophoretic separation in a 0.8% agarose gel, DNA fragments were blotted to HyBond N membrane (Amersham Biosciences, Inc.) according to the supplier's recommendations. The DNA blot was subjected to Southern hybridization using the CYP3A43 exon 1-specific oligonucleotide, 5Ј-TAC CAG GCT GGT AGC CAC AAG AAC CCA TG (Cybergene AB), end-labeled with adenosine 5Ј-[␥-32 P]triphosphate (Ͼ5000 Ci/mmol, Amersham Biosciences, Inc.). After hybridization using ExpressHyb hybridization solution (CLON-TECH), the membrane was washed twice for 30 min at room temperature with 2ϫ SSC, 0.1% SDS. This was followed by consecutive washing steps at gradually increasing stringent conditions, two washes for 30 min each at 55°C with 2ϫ, 1ϫ, and 0.5ϫ SSC. Radioactive signal was visualized either by autoradiography or by using a phosphorimaging analyzer (Fujix BAS2000).
For cloning purposes, a 1-l aliquot of the PCR reaction was subjected to a 10 cycle reamplification with 10 s at 92°C, 30 s at 60°C, and 1 min at 72°C. The PCR was performed in 50 l using 2.5 units of Taq polymerase with 2 mM MgCl 2 , 200 M dNTPs, and 120 nM of each primer.
Polyadenylated RNA was isolated by subjecting 120 g of human liver total RNA (Invitrogen) to a two round purification with Poly(A) Spin mRNA Isolation Kit (New England Biolabs). Poly(A) ϩ RNA and flow-through were precipitated with ethanol and reverse-transcribed as described for total RNA (above). Two l of the RT reaction were directly used in quantitative real-time PCR.
Testosterone Hydroxylase Activity Assay-Enzyme activity of total COS-7 cell lysates and human liver microsomes (Biopredic) was measured for 60 min at 37°C in 500 l of 0.01 M phosphate-buffered saline, pH 7.4, 0.1 mM EDTA using 100 M [4-14 C]testosterone (Amersham Biosciences, Inc.) as substrate. The reaction was initiated by the addition of 1 mM NADPH (Sigma) and terminated by extraction with ethyl acetate. After extraction, separation and evaporation of the organic phase, products were dissolved in methanol and subjected for silica gel thin layer chromatography (J.T. Baker) in dichloromethane:acetone (3:1). Radioactive products were visualized using autoradiography.

RESULTS
Intergenic Splicing between CYP3A43 and CYP3A4 Genes-Systematic analysis of cytochrome P450 2C transcripts in human epidermis and liver revealed that, in addition to the canonical CYP2C mRNAs, several chimeric molecules were also produced (10,28). These RNAs were characterized by encompassing exons from different CYP2C genes, and some also had the potential to encode for chimeric CYP2C proteins. To test the hypothesis that intergenic mRNA formation is not only restricted to the CYP2C genes, we examined whether hybrid mRNA molecules were also formed between distinct members of the CYP3A gene cluster. Specifically, we performed nested RT-PCR on human liver RNA with primers from CYP3A43 exon 1 and CYP3A4 exon 13. The amplification resulted in three DNA fragments. Cloning and DNA sequence analysis revealed that these encompassed CYP3A43 exon 1 spliced to CYP3A4 exons (Fig. 2). The exons were joined at canonical splice sites, suggesting that the hybrid mRNA molecules were formed during pre-mRNA splicing. One combination consisted of all 13 exons characteristic of the CYP3A mRNAs and thus is capable of encoding for a chimeric CYP3A protein. Worth noting is that, in the other two combinations, the CYP3A4 exons that are spliced immediately after the CYP3A43 exon 1, i.e. exon 4 and exon 7, are in the same phase as exon 2. Therefore, all three detected mRNA molecules are characterized by open reading frames that start at the canonical ATG initiation codon and terminate at the canonical TGA termination codon (Table II).
Nested RT-PCR amplification using 3A4 exon 1 forward and 3A43 exon 13 reverse primers resulted in no products (data not shown). This suggests directionality in the formation of CYP3A43/3A4 intergenic mRNAs.
To obtain additional evidence by a non-PCR-based method for the presence of intergenic CYP3A mRNAs in human liver, the RNase protection analysis was employed. Specifically, a riboprobe was generated (see "Materials and Methods") that spans 68 bases of the CYP3A43 exon 1 to CYP3A4 exon 2 splice junction. The CYP3A43 exon 1 overlap was chosen to be of sufficient length (43 bases) that would allow efficient hybridization under the conditions used; however, the CYP3A4 exon 2 overlap was minimized to only 25 bases, to destabilize formation of RNA duplexes with the highly abundant CYP3A4 mRNA. As anticipated, protected fragments of a size of 68 and 43 bases were clearly visible (Fig. 3). Moreover, the intensity of the protected fragments increased with increasing amounts of input human liver RNA. The ratio of the two protected fragments indicated that the joining of CYP3A43 exon 1 to CYP3A4 exon 2 is a much more rare event than the canonical joining of CYP3A43 exons 1 and 2.
Intergenic Splicing between CYP3A43 and CYP3A5 Genes-In similarity with the CYP3A43/3A4 chimeric mRNA molecules, CYP3A transcripts encompassing CYP3A43 and CYP3A5 exons are also formed in human liver. Specifically,

Natural trans-Splicing in Human Cells
nested RT-PCR amplification using 3A43 exon 1 forward and 3A5 exon 13 reverse primers resulted in two fragments. Cloning and subsequent sequence analysis revealed that both consisted of CYP3A43 exon 1 joined to CYP3A5 exons (Fig. 4). All exons were spliced at canonical sites, indicating that CYP3A43/ 3A5 hybrid mRNA formation was a result of a pre-mRNA splicing process. The CYP3A43/3A5 chimeric mRNAs also have directionality, as no products were detected in a nested RT-PCR experiment with CYP3A5 exon 1 sense primers combined with CYP3A43 exon 13 antisense primers (data not shown).
Lack of Intergenic Transcript between CYP3A43 and CYP3A7-In contrast to the CYP3A43/3A4 and CYP3A43/3A5 intergenic mRNA molecules, no transcripts encompassing combinations between CYP3A43 and CYP3A7 exons could be detected. Nested RT-PCR amplification using 3A43 exon 1 forward and 3A7 exon 13 reverse primers resulted in no product in human liver RNA (data not shown). Moreover, no product was detected with 3A7 exon 1 sense and 3A43 exon 13 antisense primers either (data not shown). We also repeated the RT-PCR amplification using RNA isolated from the HepG2 hepatoblastoma cell line that is characterized by higher CYP3A7 expression than adult liver. However, RT-PCR amplification with 3A43 exon 1 sense and 3A7 exon 13 antisense primers still resulted in no product (data not shown).
The Mechanism of Intergenic CYP3A Formation Is trans-Splicing-An interpretation of the mechanism that generates CYP3A43/3A4 and CYP3A43/3A5 intergenic mRNAs could be the existence of a duplicated CYP3A43 exon 1 sequence, having the same orientation as the CYP3A4 and the CYP3A5 genes, that is used in a cis-splicing process. To test this possibility, Southern hybridization was performed on human DNA digested with various restriction endonucleases. A CYP3A43 exon 1-specific oligonucleotide was used as the probe. Fig. 5 shows that hybridization resulted in single bands in all four DNA samples. Moreover, the estimated sizes for the DNA fragments fit well with those deduced from the genomic sequence, i.e. 4240, 1224, 4928, and 1961 bp for SpeI-, NdeI-, EcoRI-, and BglII-digested DNA, respectively. These results provide strong evidence that there is only a single CYP3A43 exon 1 in the human genome and thus indicate that the CYP3A43 gene is indeed the source of the 3A43 exon 1 sequences in the CYP3A43/3A4 and CYP3A43/3A5 intergenic mRNA molecules. The absence of a duplicated CYP3A43 exon 1 is also supported by the available genomic GenBank entries that cover the CYP3A locus (Fig. 1). Therefore, because the CYP3A43 gene is encoded by a different DNA strand than the other members of the CYP3A cluster, including the genes for CYP3A4 and CYP3A5, the formation of the identified intergenic CYP3A transcripts can only be rationalized by splicing between distinct pre-mRNA molecules, i.e. trans-splicing.
The Full-length Hybrid CYP3A43/3A4 Is Catalytically Active-A majority of the detected trans-spliced CYP3A mRNAs

TABLE II Correlation between translational reading frame and trans-splicing
Note that all but one of the CYP3A exons that were found to juxtapose to the CYP3A43 exon 1 have a reading frame that is identical to the reading frame of exon 2.

Exon
Reading frame Intergenic species has open reading frames that start at the canonical ATG codon and end at the canonical TGA termination codon (Table II). Therefore, the potential of coding hybrid, full-length or partial length, CYP3As exists. To address the question whether the encoded chimeric CYP3A proteins are enzymatically active, we measured testosterone 6␤-hydroxylase activity, a typical hu- Natural trans-Splicing in Human Cells 5886 man CYP3A enzymatic parameter (33), of heterologously expressed chimeric CYP3As. cDNAs encoding the two longest hybrid CYP3As detected, the full-length (1) 3A43 -(2-3-4-5-6-7-8-9-10-11-12-13) 3A4 and the partial length (1) 3A43 -(4-5-6-7-8-9-10-11-12-13) 3A4 species, were subcloned in the expression vector pcDNA1.1. Chimeric CYP3A43/3A4 proteins were transiently expressed in COS-7 cells, and sonicated total cell lysates were subjected to testosterone hydroxylase activity assay. CYP3A4 expressed in COS-7 cells and human liver microsomes served as controls for the heterologous expression and the testosterone hydroxylase reaction. As anticipated, the fulllength chimeric CYP3A43/3A4 encoded by the (1) 3A43 -(2-3-4-5-6-7-8-9-10-11-12-13) 3A4 cDNA had the capacity of hydroxylating testosterone, resulting in the formation of a product that had the same chromatographic properties as the main products of CYP3A4 or human liver microsome reactions (Fig. 7). In contrast, the shorter hybrid CYP3A43/3A4 encoded by (1) 3A43 -(4-5-6-7-8-9-10-11-12-13) 3A4 trans-spliced RNA was found inactive in this assay, as it did not result in the production of any detectable product (Fig. 7). Therefore, we concluded that the full-length trans-spliced CYP3A43/3A4 mRNA encodes a functional hybrid cytochrome P450 enzyme. DISCUSSION trans-Splicing between Distinct CYP3A Genes-The finding of CYP3A43/3A4 and CYP3A43/3A5 hybrid mRNA molecules indicates that, in similarity with recent observations of intergenic CYP2C RNAs, CYP3A expression on chromosome 7q21-22.1 also results in chimeric transcripts. This suggests that natural hybrid mRNA formation may be a generalized process and is not only limited to the CYP2C genes on chromosome 10q24. Moreover, by taking advantage of the unique arrangement of the CYP3A locus, we predict that the underlying mechanism of chimeric CYP3A43/3A4 and CYP3A43/3A5 mRNA formation is trans-splicing. Although trans-splicing is a widespread phenomenon among lower eukaryotes (34), only a few isolated cases have been reported in mammals (11,12,(35)(36)(37). A majority of these hybrid mRNA molecules contain exons from DNA segments located on different chromosomes. Therefore, the most likely mechanism for the formation of these transcripts was proposed to be trans-splicing; however, DNA rearrangements have not been conclusively ruled out.
The map of the CYP3A locus provides an indisputable example that DNA duplications may result in a variety of solitary exons or sets of exons within gene clusters (Ref. 38; Fig. 1). Duplicated exons may even be identical to the original ones of the corresponding gene (38,39). However, the genomic hybridization experiment performed, as well as the sequence of the CYP3A locus, provide solid evidence that the CYP3A43 exon 1 is a single copy exon, and thus the CYP3A43 gene is the source of that exon in the intergenic CYP3A mRNAs. Therefore, the mechanism generating the hybrid CYP3A mRNAs, containing sequences from both DNA strands, has to be trans-splicing.
The recent finding of repetitions of exons in mRNAs from the rat carnitine octanoyltransferase (20), sodium channel SNS (21), hypertension-related SA (22), and the human Sp1 (23) genes are also likely to be the result of trans-splicing. However, the CYP3A mRNAs characterized in this report represent significantly more complex products of trans-splicing, as these are not limited to single genes.
The exclusive usage of CYP3A43 exon 1 as the 5Ј exon during these trans-splicing reactions may indicate the existence of sequences in the CYP3A43 exon 1 and/or its flanking intron 1 that promote the intermolecular reaction of CYP3A pre-mRNAs. The fact that CYP3A43 intron 1 is the longest intron in the human CYP3A locus (see GenBank entries in Fig. 1), is reminiscent of an earlier finding that exon scrambling, an additional nontypical alternative splicing process, most frequently occurs with exons that are flanked by large introns (2).
trans-Spliced CYP3A mRNAs-Taking advantage of realtime PCR (31), we extended the qualitative RT-PCR data to quantitative measurements. This resulted in the finding that the abundance of the trans-spliced mRNA molecules is low, ϳ2-3 orders of magnitude less than that of the canonical CYP3A43. It must be noted, however, that the quantified (1) 3A43 -(4-5-6-7-8-9-10-11-12-13) 3A4 mRNA is not the most abundant chimeric species. The data of Fig. 2, as well as the use of the processive Expand amplification system, 2 provide clear evidence that the 13 exon-containing molecule, i.e. the (1) 3A43 -(2-3-4-5-6-7-8-9-10-11-12-13) 3A4 mRNA, is the most frequently occurring combination. This is also substantiated by the results of the RNase protection analysis (Fig. 3), which suggest that the abundance of this mRNA is about 1% that of CYP3A43.
During recent years, a large amount of data favors a model in which in vivo synthesis of mature mRNAs is the product of an "RNA factory" in which transcription, capping, splicing and polyadenylation occur in a mutually regulated way (40). However, during the formation of the CYP3A43/3A4 and the CYP3A43/3A5 trans-spliced mRNA molecules, transcription and trans-splicing are likely to occur apart from each other, both spatially and in time. The finding that the (1) 3A43 -(4-5-6-7-8-9-10-11-12-13) 3A4 CYP3A43/3A4 mRNA molecule is polya-2 C. Finta and P. G. Zaphiropoulos, unpublished observations. denylated indicates that trans-splicing does not interfere with polyadenylation. This observation is consistent with a previous finding that in vivo trans-splicing and polyadenylation may occur independently (41).
Function of the trans-Spliced CYP3As-The exact biological significance of the chimeric CYP3A mRNAs remains to be elucidated. Low abundance may suggest that these transcripts only represent "biological noise" of gene expression. However, there is a non-randomness in this process, as only combinations with the CYP3A43 exon 1 upstream of either CYP3A4 or CYP3A5 exons, but not vice versa, were detected. Moreover, there is a clear preference in the choice of the trans-spliced exons. Specifically, CYP3A43 exon 1 predominantly juxtaposes to exons having the same reading frame as exon 2; therefore, the potential for coding hybrid CYP3A proteins, full-length and deletion derivatives, exists (Table II). Nevertheless, the reason for the directionality and the biased exon choice may also reflect some, as yet unclear, structural property of the transspliced pre-mRNAs.
Interestingly, the major identified CYP3A43/3A4 mRNA encompasses all 13 exons characteristic of the canonical CYP3A transcripts, and thus has the potential to code for a full-length, functional CYP3A protein (Fig. 2). The functionality of this full-length hybrid CYP3A43/3A4 protein was experimentally confirmed by demonstrating its competence in testosterone 6␤-hydroxylation reaction. This protein only differs from the canonical CYP3A4 at the NH 2 terminus by 7 amino acids of the 24 amino acids encoded by CYP3A43 exon 1. It is generally accepted that the hydrophobic NH 2 -terminal domain is mainly responsible for targeting, insertion, and retention of cytochrome P450s in membranes of appropriate cell compartments (42). Thus, it is conceivable that the chimeric protein may have a cellular localization altered from that of CYP3A4. Moreover, recent findings revealed that, although NH 2 -terminal modifications that are frequently used to facilitate heterologous expression of CYPs, have no measurable effect on the CYP3A4 enzyme turnover rate, these do result in changes in the complex formation between the cytochrome P450 and the cytochrome P450 oxidoreductase (43). Hence, variations in the membrane binding NH 2 -terminal domain conceivably result in alterations in the in vivo metabolizing properties of the enzyme. This is consistent with our findings that 7 amino acid changes in the NH 2 -terminal region result in a significant decrease of catalytic activity (Fig. 7).
Complexity of Human Gene Expression-A recent surprising outcome of the ongoing eukaryotic genome projects is the apparent paradox between number of genes and complexity of an organism, i.e. human versus Caenorhabditis elegans or Drosophila (44,45). It is therefore conceivable that the collection of alternative spliced mRNAs, including the intergenic combinations, rather than the gene number itself, may more accurately reflect how complex an organism actually is.
Comprehensive analyses of the human transcriptome clearly indicate that the number of unique mRNA molecules may by far exceed the number of genes (46). Relevant calculations have led to the conclusion that the main underlying process, alternative splicing, is a more frequent phenomenon than previously anticipated, and thus represents "more the rule than the exception" (47). The unusually highly complicated pattern of hybrid CYP3A transcripts together with the recent findings of various chimeric CYP2C RNAs (10, 28) collectively suggest that intergenic mRNA formation may represent a generalized splicing pathway that deepens the complexity of splicing patterns in gene families. Moreover, the identified hybrid CYP3A transcripts indicate that an underlying mechanism for intergenic mRNA synthesis in mammalian cells may involve trans-splicing.