Signals for the selection of a splice site in pre-mRNA. Computer analysis of splice junction sequences and like sequences

Identification of mRNA that encodes an alternative form of H-CAM(CD44) in lymphoid and nonlymphoid tissues

The cluster of differentiation 44 (CD44), hereafter referred to as H-CAM(CD44), represents a novel class of polymorphic (Mr 80,000-215,000) cell adhesion molecules that are involved in cell-cell and cell-matrix adhesion events in a variety of organ systems. We report the detection of distinct mRNAs, in both hematopoietic and nonhematopoietic human cell lines, that encode H-CAM(CD44) with different cytoplasmic domains. Genomic Southern blot analyses indicate that the exons encoding these two cytoplasmic domains are located on the same approximately 16 kilobase (kb) Eco RI restriction fragment. Restriction endonuclease and Southern blot analyses performed on polymerase chain reaction (PCR) amplification copies of these mRNAs confirm that their sequences correspond with previously reported cDNA sequences. A consensus splice donor site which is conserved in human, baboon, and mouse mRNAs that encode a molecule with an elongated cytoplasmic domain (H-CAM-L) is utilized to generate a distinct but low-abundance mRNA species that encodes H-CAM(CD44) with a truncated cytoplasmic domain of only three amino acids (H-CAM-S). Estimations of the relative abundance of these mRNA species in B-lymphoblastoid cells using the PCR amplification technique exhibit average H-CAM-L/H-CAM-S ratios ranging between 100 and 200. Therefore, H-CAM (CD44)-mediated adhesive events may be regulated through a differential capacity of H-CAM-L and H-CAM-S to interact with the cytoskeleton and to participate in intracellular signaling events.

Exon mutations uncouple 5' splice site selection from U1 snRNA pairing

It has previously been shown that a mutation of yeast 5' splice junctions at position 5 (GUAUGU) causes aberrant pre-mRNA cleavages near the correct 5' splice site. We show here that the addition of exon mutations to an aberrant cleavage site region transforms it into a functional 5' splice site both in vivo and in vitro. The aberrant mRNAs are translated in vivo. The results suggest that the highly conserved G at the 5' end of introns is necessary for the second step of splicing. Further analyses indicate that the location of the U1 snRNA-pre-mRNA pairing is not affected by the exon mutations and that the precise 5' splice site is selected independent of this pairing.

A 62,000 molecular weight spliceosome protein crosslinks to the intron polypyrimidine tract

Incubation in HeLa nuclear extract of a 32P-labeled 61 nucleotide-long RNA corresponding to the lariat branch site/polypyrimidine tract/3' splice site of the first intron of human beta-globin pre-mRNA led to the crosslinking of a single protein of approximately 62,000 mol. wt. (p62). p62 corresponds to a polypyrimidine tract-binding protein recently described by Garcia-Blanco et al. (Genes & Dev. 3: 1874-1886, 1989). Crosslinking of p62 to the 61 nt RNA was highly sequence specific. No p62 crosslinking was observed with a 60 nt pGEM vector RNA, a 63 nt RNA antisense to the 61-mer or a 72 nt U2 RNA sequence. p62 crosslinking to the 61 nt RNA was competed by unlabeled 61 nt RNA, by beta-globin pre-mRNA containing intron 1, and by poly(U) and poly(C), but was competed to a lesser extent or not at all by pGEM RNA, a beta-globin RNA lacking intron 1, or poly(A). Experiments with mutated RNAs revealed that neither the lariat branch site adenosine nor the 3' splice site were required for p62 crosslinking to polypyrimidine tract-containing RNA. Elimination of the polypyrimidine tract reduced p62 crosslinking, as did mutation of a polypyrimidine tract UU dinucleotide to GA. However, replacement of a pyrimidine-rich tract immediately adjacent (3') to the lariat branch site with a 57% A + G pGEM vector RNA sequence also significantly reduced p62 crosslinking, indicating the involvement of both this pyrimidine-rich region and the classical polypyrimidine tract adjacent to the 3' splice site. The sites of protein interaction were further defined by RNase H protection experiments, the results of which confirmed the patterns of p62 crosslinking to mutant RNAs. p62 crosslinking was efficiently competed by a DNA oligonucleotide having the same sequence as the 61 nt RNA, showing that p62 requires neither ribose 2' OH groups nor uracil bases for its interaction with the polypyrimidine tract. p62 was not crosslinked to double-stranded 61 nt RNA. Q-Sepharose chromatography of HeLa nuclear extract yielded an unbound fraction (QU) in which p62 was the only polypyrimidine tract-crosslinkable protein and a bound fraction (QB) in which, surprisingly, several non-p62 proteins were crosslinkable to the polypyrimidine tract RNA. Yet, when the two Q-Sepharose fractions were combined, p62 strongly out-competed the otherwise-crosslinkable QB proteins for polypyrimidine tract RNA crosslinking. This indicates that p62 may have the highest affinity and/or crosslinking efficiency for the intron polypyrimidine tract of any HeLa nuclear protein.

Exon as well as intron sequences are cis-regulating elements for the mutually exclusive alternative splicing of the beta tropomyosin gene

The beta tropomyosin gene contains two internal exons which are spliced in a mutually exclusive manner. Exon 6B is specifically included in the mature transcripts expressed in skeletal muscle or cultured myotubes, while exon 6A is a myoblast- or smooth muscle-specific exon. The intron between them, which is never spliced in normal conditions, contains two characteristic features: first, the unusual location of the branch point at position -105 from the acceptor, and second, the presence of a very long pyrimidine stretch upstream of the skeletal muscle exon. In this study we designed a number of sequence modifications to investigate the role of these two elements and of a computer-predicted secondary structure in the mutually exclusive splicing of the two exons. We found that mutations in the skeletal exon as well as in the upstream intron could change in vivo the tissue-specific pattern as well as the mutually exclusive character of the two exons. Our results suggest that the unusual position of the branch point does not prevent the utilization of exon 6B in myoblasts and that the region around the acceptor site of exon 6B and the polypyrimidine tract have an important role in this control. Last, we discuss the possible implications of secondary structures.

Computer prediction of the exon-intron structure of mammalian pre-mRNAs

A novel approach to the problem of prediction of protein-coding regions is suggested. This approach combines the site prediction methods to predict splicing sites and the global coding region prediction methods to choose the best variant of spliced mRNA. One of the advantages of the suggested algorithm is that the resulting mRNA or protein sequence may then be immediately analyzed further. The true mRNA either coincides with the predicted one or ranks high in the list of variants. In the latter situation the predicted mRNA usually differs from the true one in only one or two of several exons. The combined approach allows the use of a priori information (e.g. the putative protein length or the number of exons). It is possible to use additional parameters not considered here, such as the preferred lengths of exons and introns, and particularly the preferred position of introns in the reading frame and the preferred codon position of exon termini.

Exon as well as intron sequences are cis-regulating elements for the mutually exclusive alternative splicing of the beta tropomyosin gene

The beta tropomyosin gene contains two internal exons which are spliced in a mutually exclusive manner. Exon 6B is specifically included in the mature transcripts expressed in skeletal muscle or cultured myotubes, while exon 6A is a myoblast- or smooth muscle-specific exon. The intron between them, which is never spliced in normal conditions, contains two characteristic features: first, the unusual location of the branch point at position -105 from the acceptor, and second, the presence of a very long pyrimidine stretch upstream of the skeletal muscle exon. In this study we designed a number of sequence modifications to investigate the role of these two elements and of a computer-predicted secondary structure in the mutually exclusive splicing of the two exons. We found that mutations in the skeletal exon as well as in the upstream intron could change in vivo the tissue-specific pattern as well as the mutually exclusive character of the two exons. Our results suggest that the unusual position of the branch point does not prevent the utilization of exon 6B in myoblasts and that the region around the acceptor site of exon 6B and the polypyrimidine tract have an important role in this control. Last, we discuss the possible implications of secondary structures.

Beta-globin transcripts carrying a single intron with three adjacent nucleotides of 5' exon are efficiently spliced in vitro irrespective of intron position or surrounding exon sequences

To examine the role of exon sequences and intron position in the splicing of an mRNA precursor, we prepared series of sense or anti-sense transcripts of human beta-globin cDNA in which a cassette containing the beta-globin first intron was inserted into one of seven unusual positions. The intron cassette consisted of the intron alone (ml), the intron with three adjacent base pairs of the 5' exon (MI), or the intron with both 5' and 3' exon sequences. All these transcripts were examined in an in vitro splicing system with a HeLa cell nuclear extract. The sense transcripts carrying MI cassette were spliced efficiently and independently of the intron position, except when the 3' exon was too short. The anti-sense transcripts carrying MI cassette produced significantly less spliced products than did those of the sense transcripts. This was mostly because of the instability of the anti-sense transcripts, and the actual splicing efficiency was similar to that seen in the sense transcripts. Sense or anti-sense transcripts carrying ml cassette were spliced to various extents depending on the surrounding sequences. The results indicate that only three nucleotides of the 5' exon are required as specific exon sequences in the splicing of an mRNA precursor carrying a single intron, and that the intron position does not significantly affect the splicing efficiency in vitro.

Structure and developmental expression of the D alpha 2 gene encoding a novel nicotinic acetylcholine receptor protein of Drosophila melanogaster

Nicotinic acetylcholine receptors (nAChRs) represent a heterogeneous group of excitatory neurotransmitter receptors in the insect brain. We have characterized the D alpha 2 gene of Drosophila melanogaster, a new member of the nAChR gene family. The protein coding region is interrupted by six introns. The positions of three of these introns are shared with all other nAChR genes. The deduced D alpha 2 protein shows the structural features of ligand-binding nAChR alpha-subunits. Cytogenetically, the D alpha 2 gene maps at position 96A of the 3rd chromosome, close to the ALS gene which also encodes an alpha-like nAChR subunit. D alpha 2 transcripts are predominantly expressed in late embryos and in fly heads.

Mechanism for cryptic splice site activation during pre-mRNA splicing

The 5' splice site of a pre-mRNA is recognized by U1 small nuclear ribonucleoprotein particles (snRNP) through base pairing with the 5' end of U1 small nuclear RNA (snRNA). Single-base substitutions within a 9-nucleotide 5'-splice-site sequence can abolish or attenuate use of that site and, in higher eukaryotes, can also activate nearby "cryptic" 5' splice sites. Here we show that the effects of single-base substitutions within a 5' splice site can be completely or partially suppressed by cis mutations that improve the overall complementarity of the site to U1 snRNA. We further show that in the presence of the normal 5' splice site, a cryptic 5' splice site can be activated by increasing its complementarity to U1 snRNA. U1 snRNP binding experiments confirm that cryptic 5' splice sites are activated when their affinity for U1 snRNP approaches that of the authentic 5' splice site. Based upon these results, we propose a spliceosome competition model for 5'-splice-site selection and cryptic 5'-splice-site activation. We discuss our results with regard to the factors involved in 5'-splice-site recognition.

The essential pre-mRNA splicing factor SF2 influences 5' splice site selection by activating proximal sites

SF2 is a 33 kd protein factor required for 5' splice site cleavage and lariat formation during pre-mRNA splicing in HeLa cell extracts. In addition to its essential role in constitutive splicing, SF2 can strongly influence 5' splice site selection. When pre-mRNAs containing multiple cis-competing 5' splice sites are spliced in vitro, high concentrations of purified SF2 promote the use of the 5' splice site closest to the 3' splice site. However, SF2 discriminates properly between authentic and cryptic splice sites. These effects of SF2 on splice site selection may reflect the cellular mechanisms that prevent exon skipping and ensure the accuracy of splicing. In addition, alterations in the concentration or activity of SF2, and of other general splicing factors, may serve to regulate alternative splicing in vivo.

Myosin light-chain 1/3 gene alternative splicing: cis regulation is based upon a hierarchical compatibility between splice sites

The mechanisms involved in the selective joining of appropriate 5' and 3' splice sites are still poorly understood in both constitutive and alternatively spliced genes. With two promoters associated with different exons, the myosin light-chain 1/3 gene generates two pre-mRNAs that also differ by the use of a pair of internal exons, 3 and 4, that are spliced in a mutually exclusive fashion. When the promoter upstream from exon 1 is used, only exon 4 is included. If the promoter upstream from exon 2 is used, only exon 3 is included. In an attempt to understand the molecular basis for the mutually exclusive behavior of these two exons and the basis of their specific selection, a number of minigene constructs containing exons 3 and 4 were tested in a variety of homologous or heterologous cis and trans environments. The results demonstrate that the mutually exclusive behavior of myosin light-chain exons 3 and 4 and selection between the two exons are cis regulated and are affected by the nature of the flanking sequences. Both exons competed for the common flanking 5' and 3' splice sites. Flanking exons were found that favored inclusion into mature mRNA of exon 3, exon 4, both, or neither, suggesting a specific cooperative interaction between certain 5' and 3' splice sites. Thus, alternative splicing of myosin light-chain 1/3 pre-mRNAs is regulated in cis by a hierarchy of compatibilities between pairs of 5' and 3' splice sites.

Alternative splicing of the sex determination gene transformer-2 is sex-specific in the germ line but not in the soma

The transformer-2 (tra-2) gene of Drosophila melanogaster plays essential roles in both sexual differentiation in the female soma and spermatogenesis in the male germ line. In the female soma, tra-2 is known to act with other genes in the sex determination regulatory cascade to control the sex-specific alternative splicing of transcripts from the doublesex gene. Here, we determine whether or not any sex-specific tra-2 products are expressed that may account for either of these sex-specific activities. Sequence analysis of the tra-2 gene and 10 tra-2 cDNA clones coupled with nuclease protection analysis reveals a variety of alternatively spliced tra-2 mRNAs that each encode one of four distinct but overlapping polypeptides. Three of the encoded polypeptides contain both a ribonucleoprotein consensus sequence and arginine/serine-rich regions, suggesting a direct role for these products in RNA splicing. We show that although two transcripts are expressed male specifically in the germ line, the tra-2 transcripts expressed in the soma are not sex-specific. The translation of products from a tra-2-lacZ fusion gene in both sexes suggests that the female-specific functioning of tra-2 in somatic tissues is not attributable to a translational mechanism. We suggest that tra-2 activity in somatic tissues is regulated through a post-translational sex-specific interaction with the product of the tra gene rather than through the expression of a female-specific tra-2 polypeptide.

Myosin light-chain 1/3 gene alternative splicing: cis regulation is based upon a hierarchical compatibility between splice sites

The mechanisms involved in the selective joining of appropriate 5' and 3' splice sites are still poorly understood in both constitutive and alternatively spliced genes. With two promoters associated with different exons, the myosin light-chain 1/3 gene generates two pre-mRNAs that also differ by the use of a pair of internal exons, 3 and 4, that are spliced in a mutually exclusive fashion. When the promoter upstream from exon 1 is used, only exon 4 is included. If the promoter upstream from exon 2 is used, only exon 3 is included. In an attempt to understand the molecular basis for the mutually exclusive behavior of these two exons and the basis of their specific selection, a number of minigene constructs containing exons 3 and 4 were tested in a variety of homologous or heterologous cis and trans environments. The results demonstrate that the mutually exclusive behavior of myosin light-chain exons 3 and 4 and selection between the two exons are cis regulated and are affected by the nature of the flanking sequences. Both exons competed for the common flanking 5' and 3' splice sites. Flanking exons were found that favored inclusion into mature mRNA of exon 3, exon 4, both, or neither, suggesting a specific cooperative interaction between certain 5' and 3' splice sites. Thus, alternative splicing of myosin light-chain 1/3 pre-mRNAs is regulated in cis by a hierarchy of compatibilities between pairs of 5' and 3' splice sites.

Structure, evolution and chromosomal localization of the human pregnancy-specific beta 1 glycoprotein gene family

Cloning and sequencing of the cDNA of a new member of the PS beta G gene family is reported. Comparison of the sequence with those of other PS beta G cDNAs reveals a remarkable conservation of their sequence (greater than 90%) and of their general structural organization: an NH2 domain is followed by 93- and 85-residue Ig-like domains termed A and B, respectively. Most PS beta G contain two domains A in tandem followed by a single domain B. In some PS beta G members described here, alternative splicing skips the AI domain in some transcripts, yielding two- and three-domain variants, respectively. Individual PS beta G members have specific short carboxyl domains displaying little sequence conservation. The PS beta G family is closely related to the CEA gene family. A detailed comparison of the sequence of both families is given and used to construct an evolutionary tree, using the method of Li, Wu, and Luo (1985, Mol. Biol. Evol. 2: 150-174). Computation of the number of substitutions of synonymous (Ks) and nonsynonymous (Ka) sites and of the Ks/Ka ratio suggests that the PS beta G gene family appeared concomitantly with the expansion of the placental mammals and belongs to the class of rapidly evolving genes. Very little selective pressure has been exerted on the body of the molecules, especially on domain A. The analysis also suggests that PS beta G genes encoding different carboxyl domains would have been positively selected and fixed during the evolution. The PS beta G gene family was assigned to chromosome 19, which also carries the CEA genes.

A putative ATP binding protein influences the fidelity of branchpoint recognition in yeast splicing

We previously described a dominant suppressor of the splicing defect conferred by an A----C intron branchpoint mutation in S. cerevisiae. Suppression occurs by increasing the frequency with which the mutant branchpoint is utilized. We have now cloned the genomic region encoding the prp16-1 suppressor function and have demonstrated that PRP16 is essential for viability. A 1071 amino acid open reading frame contains sequence motifs characteristic of an NTP binding fold and further similarities to a superfamily of proteins that includes members with demonstrated RNA-dependent ATPase activity. A single nucleotide change necessary to confer the prp16-1 suppressor phenotype results in a Tyr----Asp substitution near the "A site" consensus for NTP binding proteins. We propose that PRP16 is an excellent candidate for mediating one of the many ATP-requiring steps of spliceosome assembly and that accuracy of branchpoint recognition may be coupled to ATP binding and/or hydrolysis.

Isolation of the chicken middle-molecular weight neurofilament (NF-M) gene and characterization of its promoter

We have isolated and sequenced genomic DNA clones covering the coding region of the chicken mid-size neurofilament (NF-M) gene and greater than 1 kb of its 5' upstream region. The NF-M gene contains two introns which both are located within the highly conserved C-terminal region of the rod domain. The 5' end of the corresponding mRNA was assigned to a G residue 40 nucleotides upstream of the translation start site and in appropriate distance from a potential TATA box. To functionally analyze the NF-M promoter, constructs carrying 112, 222, and 1026 nucleotides of the 5' upstream region in front of a luciferase reporter gene were tested for their capability to direct luciferase expression after transient transfection into various cell lines. Significant luciferase activity was recorded both in rat phaeochromocytoma (PC12) cells and murine fibroblasts. In PC12 cells, in which neurite outgrowth is induced by nerve growth factor (NGF), expression was stimulated up to 13-fold within 3 days of NGF treatment. This closely resembles expression of the endogenous NF-M gene in response to this hormone.

Exon definition may facilitate splice site selection in RNAs with multiple exons

Interactions at the 3' end of the intron initiate spliceosome assembly and splice site selection in vertebrate pre-mRNAs. Multiple factors, including U1 small nuclear ribonucleoproteins (snRNPs), are involved in initial recognition at the 3' end of the intron. Experiments were designed to test the possibility that U1 snRNP interaction at the 3' end of the intron during early assembly functions to recognize and define the downstream exon and its resident 5' splice site. Splicing precursor RNAs constructed to have elongated second exons lacking 5' splice sites were deficient in spliceosome assembly and splicing activity in vitro. Similar substrates including a 5' splice site at the end of exon 2 assembled and spliced normally as long as the second exon was less than 300 nucleotides long. U2 snRNPs were required for protection of the 5' splice site terminating exon 2, suggesting direct communication during early assembly between factors binding the 3' and 5' splice sites bordering an exon. We suggest that exons are recognized and defined as units during early assembly by binding of factors to the 3' end of the intron, followed by a search for a downstream 5' splice site. In this view, only the presence of both a 3' and a 5' splice site in the correct orientation and within 300 nucleotides of one another will stable exon complexes be formed. Concerted recognition of exons may help explain the 300-nucleotide-length maximum of vertebrate internal exons, the mechanism whereby the splicing machinery ignores cryptic sites within introns, the mechanism whereby exon skipping is normally avoided, and the phenotypes of 5' splice site mutations that inhibit splicing of neighboring introns.

Hierarchy for 5' splice site preference determined in vivo

The relationship between preferences among alternative 5' splice sites and their sequences has been investigated for 37 sequences by assessing their use in splicing relative to the 5' splice site of IVS-2 of rabbit beta-globin. There are strong correlations between the intrinsic strength of a 5' splice site and both the extent to which it resembles the consensus sequence and the calculated stability of its interactions with U1 small nuclear RNA. However, present methods of calculating either of the latter values do not allow predictions to be made of the relative preferences among a small number of sequences.

In vitro splicing of mutually exclusive exons from the chicken beta-tropomyosin gene: role of the branch point location and very long pyrimidine stretch

The chicken beta-tropomyosin gene contains 11 exons, two of which are spliced into mRNA only in skeletal muscle. One pair of alternative exons, 6A and 6B, is found in the middle of the gene; they are spliced in a mutually exclusive manner. The non-muscle splice 6A-7 is by far the predominant in vitro reaction in a HeLa cell nuclear extract. A minor product is the 6A-6B splice, which is excluded in all tissues. This minor product results from the use of a branch point located 105 nt upstream of the 3' end of the intron separating exons 6A and 6B. The region between the branch point sequence and the final AG contains a stretch of approximately 80 pyrimidines. We have examined the role of the distance of the branchpoint to the 3' splice site and of the sequences between these two elements. Our results suggest that at least two cis-acting elements contribute to the mutual exclusivity of exons 6A and 6B. The intron between exons 6A and 6B is intrinsically poorly 'spliceable' both because the branch point is too far upstream of the 3' end of the intron to give efficient splicing and because of the particular sequence lying between this branch point and the 3' splice site.

Exon definition may facilitate splice site selection in RNAs with multiple exons

Interactions at the 3' end of the intron initiate spliceosome assembly and splice site selection in vertebrate pre-mRNAs. Multiple factors, including U1 small nuclear ribonucleoproteins (snRNPs), are involved in initial recognition at the 3' end of the intron. Experiments were designed to test the possibility that U1 snRNP interaction at the 3' end of the intron during early assembly functions to recognize and define the downstream exon and its resident 5' splice site. Splicing precursor RNAs constructed to have elongated second exons lacking 5' splice sites were deficient in spliceosome assembly and splicing activity in vitro. Similar substrates including a 5' splice site at the end of exon 2 assembled and spliced normally as long as the second exon was less than 300 nucleotides long. U2 snRNPs were required for protection of the 5' splice site terminating exon 2, suggesting direct communication during early assembly between factors binding the 3' and 5' splice sites bordering an exon. We suggest that exons are recognized and defined as units during early assembly by binding of factors to the 3' end of the intron, followed by a search for a downstream 5' splice site. In this view, only the presence of both a 3' and a 5' splice site in the correct orientation and within 300 nucleotides of one another will stable exon complexes be formed. Concerted recognition of exons may help explain the 300-nucleotide-length maximum of vertebrate internal exons, the mechanism whereby the splicing machinery ignores cryptic sites within introns, the mechanism whereby exon skipping is normally avoided, and the phenotypes of 5' splice site mutations that inhibit splicing of neighboring introns.

Scanning from an independently specified branch point defines the 3' splice site of mammalian introns

During pre-messenger RNA splicing the lariat branch point in mammalian introns is specified independently of the 3' splice site by the sequence surrounding the branch point and by an adjacent downstream polypyrimidine tract. The 3' splice site is dispensable for spliceosome assembly and cleavage at the 5' splice site, and is itself determined by a scanning process that recognizes the first AG located 3' of the branch point/polypyrimidine tract, irrespective of distance or sequence environment.

Identification of functional U1 snRNA-pre-mRNA complexes committed to spliceosome assembly and splicing

Although both U1 and U2 snRNPs have been implicated in the splicing process, their respective roles in the earliest stages of intron recognition and spliceosome assembly are uncertain. To address this issue, we developed a new strategy to prepare snRNP-depleted splicing extracts using Saccharomyces cerevisiae cells conditionally expressing U1 or U2 snRNP. Complementation analyses and chase experiments show that a stable complex, committed to the splicing pathway, forms in the absence of U2 snRNP. U1 snRNP and a substrate containing both a 5' splice site and a branchpoint sequence are required for optimal formation of this commitment complex. We developed new gel electrophoresis conditions to identify these committed complexes and to show that they contain U1 snRNA. Chase experiments demonstrated that these complexes are functional intermediates in spliceosome assembly and splicing. Our results have implications for the process of splice site selection.

Structure of bovine immunoglobulin constant region heavy chain gamma 1 and gamma 2 genes

Two bovine immunoglobulin constant region gamma heavy chain germline gene sequences are described. A gamma 1 gene was cloned from a lambda 2001 calf liver library screened with a human gamma 4 (pBRH4.1) probe and is contained in a 5.8 kb BamH1 hybridizing fragment. The gamma 2 gene was from an EMBL4 lambda library and is in a 6.6 kb BamH1 fragment. Each of these genes is arranged in four exons corresponding to the three CH domains and the hinge of gamma heavy chain genes; normal RNA splice and polyadenylation sites are present. The translated C-terminal peptide sequences of the genes match exactly the equivalent peptide sequences of bovine IgG1 and IgG2 heavy chains, identifying them as gamma 1 and gamma 2. The derived protein sequences reveal 96, 80 and 83% identity of amino acid residues between their CH1, CH2 and CH3 domains. Two adjacent cysteine residues encoded in the CH1 exons suggest that, as in rabbit gamma, an extra intra-chain disulphide bond occurs in the bovine gamma heavy chains. Significant DNA rearrangement in the hinge-CH2 region is evident in the bovine gamma 2 gene, with resultant deletion and substitution of amino acid residues in the lower hinge and N-terminal portion of the CH2 domain. The Fab-Fc interface of bovine IgG2 is predicted to be sterically blocked, relative to IgG1, which has implications for effector differences between the bovine gamma subclasses.

Expression of the H-ras proto-oncogene is controlled by alternative splicing

We previously demonstrated that a point mutation in the last intron of the human H-ras oncogene causes a significant increase in its expression and transforming efficiency. Here we establish the basis of this phenomenon. Using gene reconstruction experiments, we have identified a negative-acting element in the intron that is completely inactivated by the mutation. The effects of other nucleotide alterations introduced into this region suggested that the negative element might constitute an alternative exon. Transcripts containing this putative exon were identified and S1 nuclease analysis confirmed that the mutation prevents their synthesis. The abundance of these transcripts is low, apparently due to message instability and/or defective processing. The predicted product of the alternative transcript is suggested to lack transforming potential. Our findings demonstrate that alternative splicing normally operates to suppress p21H-ras expression and that this negative control is abolished by a variety of mutations that interfere with this process.

Statistical analysis of mammalian pre-mRNA splicing sites

222 donor and 222 acceptor (including 206 pairs) non-homologous splicing sites were studied. Well known features of these were confirmed and some novel observations were made. It is (1) cCAGGGag signal in (-60)-(-58) region of acceptor sites; (2) strong complementarity between regions (-69)-(-55) and (-36)-(-22) of some of the acceptor sites, and (3) small but statistically significant correlation between discrimination energies of corresponding donor and acceptor sites.

The multiple codes of nucleotide sequences

Nucleotide sequences carry genetic information of many different kinds, not just instructions for protein synthesis (triplet code). Several codes of nucleotide sequences are discussed including: (1) the translation framing code, responsible for correct triplet counting by the ribosome during protein synthesis; (2) the chromatin code, which provides instructions on appropriate placement of nucleosomes along the DNA molecules and their spatial arrangement; (3) a putative loop code for single-stranded RNA-protein interactions. The codes are degenerate and corresponding messages are not only interspersed but actually overlap, so that some nucleotides belong to several messages simultaneously. Tandemly repeated sequences frequently considered as functionless "junk" are found to be grouped into certain classes of repeat unit lengths. This indicates some functional involvement of these sequences. A hypothesis is formulated according to which the tandem repeats are given the role of weak enhancer-silencers that modulate, in a copy number-dependent way, the expression of proximal genes. Fast amplification and elimination of the repeats provides an attractive mechanism of species adaptation to a rapidly changing environment.

Identification of members of the P-glycoprotein multigene family

Overproduction of P-glycoprotein is intimately associated with multidrug resistance. This protein appears to be encoded by a multigene family. Thus, differential expression of different members of this family may contribute to the complexity of the multidrug resistance phenotype. Three lambda genomic clones isolated from a hamster genomic library represent different members of the hamster P-glycoprotein gene family. Using a highly conserved exon probe, we found that the hamster P-glycoprotein gene family consists of three genes. We also found that the P-glycoprotein gene family consists of three genes in mice but has only two genes in humans and rhesus monkeys. The hamster P-glycoprotein genes have similar exon-intron organizations within the 3' region encoding the cytoplasmic domains. We propose that the hamster P-glycoprotein gene family arose from gene duplication. The hamster pgp1 and pgp2 genes appear to be more closely related to each other than either gene is to the pgp3 gene. We speculate that the hamster pgp1 and pgp2 genes arose from a recent gene duplication event and that primates did not undergo this duplication and therefore contain only two P-glycoprotein genes.

Genomic cloning and heterologous expression of human differentiation-stimulating factor

Differentiation-stimulating factor (D-factor) purified from mouse Ehrlich ascites cells was sequenced partially and found to be almost identical to leukemia inhibitory factor (LIF) from mouse Krebs II ascites cells. In comparison to LIF, D-factor had an additional amino-terminal serine residue. Using synthetic oligonucleotide probes designed from the murine D-factor sequence, we cloned the human gene encoding D-factor. A partial D-factor cDNA was cloned from COS-1 cells transfected with the human D-factor gene under the control of a heterologous promoter. We used this cDNA to construct a vector for direct expression of the protein in Escherichia coli. A mammalian cell expression vector was constructed using the signal sequence of interferon-alpha A linked to the D-factor cDNA. Both forms of recombinant human D-factor were active on the murine myeloid leukemia cell line M1 in a dose- and time-dependent manner for the inhibition of [3H]thymidine incorporation, and also induced phagocytosis, Fc receptor expression, and prostaglandin E2 synthesis by M1 cells.

Evidence for nuclear factors involved in recognition of 5' splice sites

To investigate soluble factors involved in pre-messenger RNA splicing we have fractionated nuclear extract by simple centrifugation to produce a supernatant pellet pair. Factors larger than 15S including U2, U4, U5, and U6 snRNPs fractionate with the pellet; U1 snRNPs distribute equally in pellet and supernatant. Each fraction is individually incompetent for splicing and spliceosome assembly; mixing restores wild type activity and assembly. The pellet fraction directs an aberrant assembly pathway in which proper 3', but improper 5' splice site recognition occurs. Complexes formed with the pellet fraction are distinguishable from wild-type complexes using native gel electrophoresis. Pellet complexes contain U1 snRNP antigens and their formation requires ATP, U1 snRNPs, U2 snRNPs, and sequences at the 3' end of the intron - properties shared with the initial steps of normal assembly and directed by sequences at the 3' end of the intron. In contrast, pellet complex assembly shows no dependence on the presence of a 5' splice junction within precursor RNA. Furthermore, binding of factors to the 5' splice junction is deficient in pellet assemblies. Thus, the pellet lacks a factor required for proper recognition of 5' splice sites. This factor can be supplied by the supernatant. Complementation occurs when supernatant U1 RNA is destroyed, suggesting that the supernatant factor recognizing 5' splice sites is not U1 snRNPs.

Complete structure of the gene for human keratin 18

The complete sequence of the human keratin 18 (K18) gene was determined. The K18 gene is 3791 bp in length and the K18 protein is coded for by seven exons. The exon structure of K18 has been conserved compared to that of other keratin genes, with the exception of a single 3' terminal exon that codes for the tail domain of the protein that is represented by two exons in epidermal keratins. The K18 gene contains an unusual AG/GC donor splice site of intron 3 instead of the consensus AG/GT sequence. This variation is not seen in any other intermediate filament genes. The promoter region of the gene contains a TATA box, six potential SP1 binding sites, and 10 copies of CACCC boxes but lacks any CCAAT boxes and is surprisingly different from the immediately 5' flanking region of the homologous mouse Endo B gene. However, both genes contain small CpG islands surrounding the 5' end of exon 1 and, in addition, conserve repetitive Alu potential transcription units approximately 300 nt upstream of the transcriptional start site.

Drosophila doublesex gene controls somatic sexual differentiation by producing alternatively spliced mRNAs encoding related sex-specific polypeptides

The doublesex (dsx) gene regulates somatic sexual differentiation in both sexes in D. melanogaster. Two functional products are encoded by dsx: one product is expressed in females and represses male differentiation, and the other is expressed in males and represses female differentiation. We have determined that the dsx gene is transcribed to produce a common primary transcript that is alternatively spliced and polyadenylated to yield male- and female-specific mRNAs. These sex-specific mRNAs share a common 5' end and three common exons, but possess alternative sex-specific 3' exons, thus encoding polypeptides with a common amino-terminal sequence but sex-specific carboxyl termini. Genetic and molecular data suggest that sequences including and adjacent to the female-specific splice acceptor site play an important role in the regulation of dsx expression by the transformer and transformer-2 loci.

A general model for the evolution of nuclear pre-mRNA introns

We present an overview of the evolution of eukaryotic split gene structure and pre-mRNA splicing mechanisms. We have drawn together several seemingly conflicting ideas and we show that they can all be incorporated in a single unified theory of intron evolution. The resulting model is consistent with the notion that introns, as a class, are very ancient, having originated in the "RNA world"; it also supports the concept that introns may have played a crucial role in the construction of many eukaryotic genes and it accommodates the idea that introns are related to mobile insertion elements. Our conclusion is that introns could have a profound effect on the course of eukaryotic gene evolution, but that the origin and maintenance of intron sequences depends, largely, on natural selection acting on the intron sequences themselves.

Identification of members of the P-glycoprotein multigene family

Overproduction of P-glycoprotein is intimately associated with multidrug resistance. This protein appears to be encoded by a multigene family. Thus, differential expression of different members of this family may contribute to the complexity of the multidrug resistance phenotype. Three lambda genomic clones isolated from a hamster genomic library represent different members of the hamster P-glycoprotein gene family. Using a highly conserved exon probe, we found that the hamster P-glycoprotein gene family consists of three genes. We also found that the P-glycoprotein gene family consists of three genes in mice but has only two genes in humans and rhesus monkeys. The hamster P-glycoprotein genes have similar exon-intron organizations within the 3' region encoding the cytoplasmic domains. We propose that the hamster P-glycoprotein gene family arose from gene duplication. The hamster pgp1 and pgp2 genes appear to be more closely related to each other than either gene is to the pgp3 gene. We speculate that the hamster pgp1 and pgp2 genes arose from a recent gene duplication event and that primates did not undergo this duplication and therefore contain only two P-glycoprotein genes.

Evidence for a class of very small introns in the gene for hypoxanthine-guanine phosphoribosyltransferase in Schistosoma mansoni

The single copy gene for the hypoxanthine-guanine phosphoribosyltransferase (HGPRTase) of the parasitic trematode, Schistosoma mansoni, contains seven introns, the first four of which are only 31, 33, 42, and 32 bases in length. These are the smallest introns ever discovered in a non-viral nuclear gene coding for protein. These very small introns possess the canonical GT...AG splice site sequences but lack the branching sequence, the secondary structure, and the minimum size of approximately 50 bases believed to be required for the splicing of eucaryotic mRNA precursors. Evidently, a somewhat different splicing mechanism for the transcripts of these very small introns is necessary. Their discovery within the genes of helminths raises theoretical considerations for the evolution of introns in eucaryotes.

A 5' splice-region G----C mutation in exon 1 of the human beta-globin gene inhibits pre-mRNA splicing: a mechanism for beta+-thalassemia

We have characterized a Mediterranean beta-thalassemia allele containing a sequence change at codon 30 that alters both beta-globin pre-mRNA splicing and the structure of the hemoglobin product. Presumably, this G----C transversion at position -1 of intron 1 reduces severely the utilization of the normal 5' splice site since the level of the Arg----Thr mutant hemoglobin (designated hemoglobin Kairouan) found in the erythrocytes of the patient is very low (2% of total hemoglobin). Since no natural mutations of the guanine located at position -1 of the CAG/GTAAGT consensus sequence had been isolated previously, we investigated the role of this nucleotide in the constitution of an active 5' splice site by studying the splicing of the pre-mRNA in cell-free extracts. We demonstrate that correct splicing of the mutant pre-mRNA is 98% inhibited. Our results provide further insights into the mechanisms of pre-mRNA maturation by revealing that the last residue of the exon plays a role at least equivalent to that of the intron residue at position +5.

Clinical and genetic characterization of nonclassic 21-hydroxylase deficiency

Nonclassic steroid 21-hydroxylase deficiency is a frequent, relatively mild disorder of cortisol biosynthesis characterized by variable signs of postnatal androgen excess. It is inherited as an allelic variant of the CYP21B gene encoding the 21-hydroxylase enzyme. CYP21B is located in the HLA histocompatibility complex, and a nonclassic allele is often associated with characteristic HLA antigens: B14;DR1. A CYP21B gene from a HLA-B14;DR1 homozygous patient with nonclassic 21-hydroxylase deficiency was cloned and analyzed. Five deviations from the normal sequence of CYP21B were found, but only one appeared likely to affect the functional integrity of the protein: codon 281, GTG, encoding valine, was changed to TTG, leucine. An oligonucleotide probe was constructed corresponding to the mutant sequence surrounding codon 281 and hybridized with DNA samples digested with restriction endonuclease Taq I. Samples from 8 nonclassic 21-hydroxylase deficiency patients carrying HLA-B14;DR1 contained a hybridizing fragment 3700 base-pairs long, indicating presence of the val-281 mutation in the CYP21B gene. In contrast, unaffected individuals and one patient who lacked HLA-B14;DR1 showed no evidence of the val-281 mutation in CYP21B. We conclude that the codon 281 mutation is a consistent molecular genetic marker for nonclassic 21-hydroxylase deficiency associated with HLA-B14;DR1.

Short donor site sequences inserted within the intron of beta-globin pre-mRNA serve for splicing in vitro

We constructed SP6-human beta-globin derivative plasmids that included possible donor site (5' splice site) sequences at a specified position within the first intron. The runoff transcripts from these templates truncated in the second exon were examined for splicing in a nuclear extract from HeLa cells. In addition to the products from the authentic donor site, a corresponding set of novel products from the inserted, alternative donor site was generated. Thus, a short sequence inserted within an intron can be an active donor site signal in the presence of an authentic donor site. The active donor site sequences included a 9-nucleotide consensus sequence, 14- or 16-nucleotide sequences at the human beta-globin first or second donor, and those at simian virus 40 large T antigen or small t antigen donor. These included 3 to 8 nucleotides of an exon and 6 to 8 nucleotides of an intron. The activity of the inserted donor site relative to that of the authentic donor site depended on the donor sequence inserted. The relative activity also strongly depended on the concentrations of both KCl (40 to 100 mM) and MgCl2 (1.6 to 6.4 mM). At the higher KCl concentrations tested, all the inserted, or proximate, donor sites were more efficiently used. Under several conditions, some inserted donor sites were more active than was the authentic donor site. Our system provides an in vitro assay for donor site activity of a sequence to be tested.

Complementary DNA cloning and sequencing of rat ovarian basic fibroblast growth factor and tissue distribution study of its mRNA

Three cDNA clones encoding rat basic fibroblast growth factor (FGF) were isolated from 10(6) independent clones prepared from a pregnant mare serum gonadotropin (PMSG)-stimulated rat ovarian cDNA library. One of the cDNA clones contained the entire coding sequence for basic FGF. The other two possessed the sequence coding the carboxy terminal 61 amino acids of rat basic FGF, the putative upstream intron sequence, and a 3'-noncoding region. The cDNAs encoding rat basic FGF predict a molecule consisting of 154 amino acid residues, which is one amino acid shorter than the human and bovine basic FGF. Otherwise, there are only 5 conservative amino acid substitutions between the rat and the human/bovine sequences. Poly A+ RNA from brain cortex and hypothalamus show a single 6.0 kb band that hybridizes to the cloned cDNA probe by Northern analyses. The observation that basic FGF mRNA is below the limits of detection in adrenal, spleen, heart, lung, kidney, liver, stomach, small intestine, large intestine, testis, and ovary support the notion that the that the high levels of the protein found in these tissues is due to storage of the mitogen in the extracellular matrix and not continuous gene expression. The significance of the abundance of mRNA in tissues which are not undergoing either active angiogenesis or cell proliferation (hypothalamus and brain cortex) is unclear but emphasizes the potential neuronotrophic function of basic FGF.

Tandem linkage of human CSF-1 receptor (c-fms) and PDGF receptor genes

A 5' untranslated exon of the human CSF-1 receptor gene (c-fms) is separated by a 26 kb intron from the 32 kb receptor coding sequences. Nucleotide sequence analysis of cloned genomic DNA revealed that the 3' end of the PDGF receptor gene is located less than 0.5 kb upstream from this exon. Similarities in chromosomal localization, organization, and encoded amino acid sequences suggest that the genes encoding the CSF-1 and PDGF receptors arose through duplication. The as yet unidentified c-fms promoter/enhancer sequences may be confined to the nucleotides separating the two genes or could potentially lie within the PDGF receptor gene itself.

Short donor site sequences inserted within the intron of beta-globin pre-mRNA serve for splicing in vitro

We constructed SP6-human beta-globin derivative plasmids that included possible donor site (5' splice site) sequences at a specified position within the first intron. The runoff transcripts from these templates truncated in the second exon were examined for splicing in a nuclear extract from HeLa cells. In addition to the products from the authentic donor site, a corresponding set of novel products from the inserted, alternative donor site was generated. Thus, a short sequence inserted within an intron can be an active donor site signal in the presence of an authentic donor site. The active donor site sequences included a 9-nucleotide consensus sequence, 14- or 16-nucleotide sequences at the human beta-globin first or second donor, and those at simian virus 40 large T antigen or small t antigen donor. These included 3 to 8 nucleotides of an exon and 6 to 8 nucleotides of an intron. The activity of the inserted donor site relative to that of the authentic donor site depended on the donor sequence inserted. The relative activity also strongly depended on the concentrations of both KCl (40 to 100 mM) and MgCl2 (1.6 to 6.4 mM). At the higher KCl concentrations tested, all the inserted, or proximate, donor sites were more efficiently used. Under several conditions, some inserted donor sites were more active than was the authentic donor site. Our system provides an in vitro assay for donor site activity of a sequence to be tested.

Comparison of in vitro and in vivo splice site selection in kappa-immunoglobulin precursor mRNA

The processing of a number of kappa-immunoglobulin primary mRNA (pre-mRNA) constructs has been examined both in vitro and in vivo. When a kappa-immunoglobulin pre-mRNA containing multiple J segment splice sites is processed in vitro, the splice sites are used with equal frequency. The presence of signal exon, S-V intron, or variable (V) region has no effect on splice site selection in vitro. Nuclear extracts prepared from a lymphoid cell line do not restore correct splice site selection. Splice site selection in vitro can be altered by changing the position or sequence of J splice donor sites. These results differ from the processing of similar pre-mRNAs expressed in vivo by transient transfection. The 5'-most J splice donor site was exclusively selected in vivo, even in nonlymphoid cells, and even in transcripts where in vitro splicing favored a 3' J splice site. The in vitro results are consistent with a model proposing that splice site selection is influenced by splice site strength and proximity; however, our in vivo results demonstrate a number of discrepancies with such a model and suggest that splice site selection may be coupled to transcription or a higher-order nuclear structure.

Cloning of the human keratin 18 gene and its expression in nonepithelial mouse cells

Human keratin 18 (K18) and the homologous mouse protein, Endo B, are intermediate filament subunits of the type I keratin class. Both are expressed in many simple epithelial cell types including trophoblasts, the first differentiated cell type to appear during mouse embryogenesis. The K18 gene was identified and cloned from among the 15 to 20 similar sequences identified within the human genome. The identity of the cloned gene was confirmed by comparing the sequence of the first two exons to the K18 cDNA sequence and transfecting the gene into various murine cell lines and verifying the encoded protein as K18 by immunoprecipitation and partial peptide mapping. The transfected K18 gene was expressed in mouse HR9 parietal endodermal cells and mouse fibroblasts even though the fibroblasts fail to express endogenous Endo B. S1 nuclease protection analysis indicated that mRNA synthesized from the transfected K18 gene is initiated at the same position as authentic K18 mRNA found in both BeWo trophoblastoma cells and HeLa cells. Pulse-chase experiments indicated that the human K18 protein is stable in murine parietal endodermal cells (HR9) which express EndoA, a complementary mouse type II keratin. Surprisingly, however, K18 was degraded when synthesized in cells which lack a type II keratin. This turnover of K18 may be an important mechanism by which epithelial cells maintain equal molar amounts of both type I and II keratins. In addition, the levels of the endogenous type I Endo B in parietal endodermal cells were compensatingly down regulated in the presence of the K18 protein, while the levels of the endogenous type II Endo A were not affected in any of the transfected cell lines.

Molecular genetic analysis of nonclassic steroid 21-hydroxylase deficiency associated with HLA-B14,DR1

Nonclassic steroid 21-hydroxylase deficiency is a frequent, relatively mild disorder of cortisol biosynthesis, characterized by variable signs of postnatal androgen excess. It is inherited as an allelic variant of the gene designated CYP21B, which encodes 21-hydroxylase. CYP21B is located in the HLA histocompatibility complex, and a "nonclassic" allelic variant is often associated with characteristic HLA antigens--B14,DR1. We cloned and analyzed the CYP21B gene from a patient homozygous for HLA-B14,DR1 who had nonclassic 21-hydroxylase deficiency. Five deviations from the normal genetic sequence of CYP21B were found, but only one appeared likely to affect the functional integrity of the protein: codon 281, GTG, encoding valine, was changed to TTG, leucine. We constructed an oligonucleotide probe corresponding to the mutant DNA sequence surrounding codon 281 and hybridized the probe with DNA samples digested with the restriction endonuclease Taql. Samples from eight patients with nonclassic 21-hydroxylase deficiency who had the haplotype HLA-B14,DR1 contained a hybridizing fragment 3700 base pairs long, indicating the presence of the valine-281 mutation in the CYP21B gene. In contrast, unaffected subjects and one patient with nonclassic deficiency who did not have HLA-B14,DR1 had no evidence of this mutation. We conclude that the mutation in codon 281 is a consistent molecular genetic marker for nonclassic 21-hydroxylase deficiency associated with HLA-B14,DR1.

Comparison of in vitro and in vivo splice site selection in kappa-immunoglobulin precursor mRNA

The processing of a number of kappa-immunoglobulin primary mRNA (pre-mRNA) constructs has been examined both in vitro and in vivo. When a kappa-immunoglobulin pre-mRNA containing multiple J segment splice sites is processed in vitro, the splice sites are used with equal frequency. The presence of signal exon, S-V intron, or variable (V) region has no effect on splice site selection in vitro. Nuclear extracts prepared from a lymphoid cell line do not restore correct splice site selection. Splice site selection in vitro can be altered by changing the position or sequence of J splice donor sites. These results differ from the processing of similar pre-mRNAs expressed in vivo by transient transfection. The 5'-most J splice donor site was exclusively selected in vivo, even in nonlymphoid cells, and even in transcripts where in vitro splicing favored a 3' J splice site. The in vitro results are consistent with a model proposing that splice site selection is influenced by splice site strength and proximity; however, our in vivo results demonstrate a number of discrepancies with such a model and suggest that splice site selection may be coupled to transcription or a higher-order nuclear structure.

Identification of the gene coding for the Endo B murine cytokeratin and its methylated, stable inactive state in mouse nonepithelial cells

The Endo B type-I keratin intermediate filament protein is first expressed at the 4- to 8-cell stage of mouse development. In the adult, its expression is restricted to a variety of simple epithelial cell types. To investigate the mechanisms responsible for the restricted expression of Endo B, the gene coding for Endo B has been identified from among the five different Endo B genes found in the mouse genome by Southern hybridization analysis and cloning all or part of four of the genes. Nuclear run-on experiments demonstrate that Endo B expression is regulated at the level of transcription. The 5' end of the active gene, designated Endo beta 1, was found to be highly methylated and in a relatively nuclease-resistant chromatin conformation in fibroblasts and myoblasts that do not express Endo B, but undermethylated and relatively sensitive to nuclease digestion in endodermal cells or F9 embryonal carcinoma cells. The inactive state of the Endo B beta 1 gene in fibroblast appears to be very stable, because somatic cell hybrids formed by the fusion of HeLa cells, which express the homologous human protein, keratin 18, and mouse fibroblasts, continue to express keratin 18 but do not activate Endo B expression. Similarly, the fusion of mouse endodermal cells and fibroblasts results in hybrids that do not extinguish Endo B expression. These results suggest that Endo B transcription is limited by two different mechanisms. In somatic cells such as fibroblasts or myoblasts, expression may be restricted by methylation and a stable, nonpermissive transcriptional state. However, in embryonal carcinoma cells, the Endo B beta 1 gene is undermethylated and in a relatively nuclease-sensitive conformation, but it is restricted by an additional, negative regulatory mechanism.

Cloning of the human keratin 18 gene and its expression in nonepithelial mouse cells

Human keratin 18 (K18) and the homologous mouse protein, Endo B, are intermediate filament subunits of the type I keratin class. Both are expressed in many simple epithelial cell types including trophoblasts, the first differentiated cell type to appear during mouse embryogenesis. The K18 gene was identified and cloned from among the 15 to 20 similar sequences identified within the human genome. The identity of the cloned gene was confirmed by comparing the sequence of the first two exons to the K18 cDNA sequence and transfecting the gene into various murine cell lines and verifying the encoded protein as K18 by immunoprecipitation and partial peptide mapping. The transfected K18 gene was expressed in mouse HR9 parietal endodermal cells and mouse fibroblasts even though the fibroblasts fail to express endogenous Endo B. S1 nuclease protection analysis indicated that mRNA synthesized from the transfected K18 gene is initiated at the same position as authentic K18 mRNA found in both BeWo trophoblastoma cells and HeLa cells. Pulse-chase experiments indicated that the human K18 protein is stable in murine parietal endodermal cells (HR9) which express EndoA, a complementary mouse type II keratin. Surprisingly, however, K18 was degraded when synthesized in cells which lack a type II keratin. This turnover of K18 may be an important mechanism by which epithelial cells maintain equal molar amounts of both type I and II keratins. In addition, the levels of the endogenous type I Endo B in parietal endodermal cells were compensatingly down regulated in the presence of the K18 protein, while the levels of the endogenous type II Endo A were not affected in any of the transfected cell lines.

Splice site selection and ribonucleoprotein complex assembly during in vitro pre-mRNA splicing

To study the determinants of splice site selection, we have inserted synthetic 5' and 3' splice sites at different positions within beta-globin genes and analyzed the resultant RNA substrates for in vitro splicing, factor binding, and complex assembly. We show that consensus 5' and 3' splice site sequences are insufficient to determine splice site utilization; in the presence or absence of the authentic site, the synthetic sites are variably active in a position-dependent manner. However, regardless of position or utilization, the synthetic 5' and 3' splice sites are bound by the appropriate splicing factors. Thus, binding of splicing factors is necessary but not sufficient for splice site utilization. Finally, we demonstrate that a block to efficient splicing can occur at multiple steps in the pathway of normal splicing complex assembly.