Current topics in genome evolution: molecular mechanisms of new gene formation.

Comparative genome analyses reveal that most functional domains of human genes have homologs in widely divergent species. These shared functional domains, however, are differentially shuffled among evolutionary lineages to produce an increasing number of domain architectures. Combined with duplication and adaptive evolution, domain shuffling is responsible for the great phenotypic complexity of higher eukaryotes. Although the domain-shuffling hypothesis is generally accepted, determining the molecular mechanisms that lead to domain shuffling and novel gene creation has been challenging, as sequence features accompanying the formation of known genes have been obscured by accumulated mutations. The growing availability of genome sequences and EST databases allows us to study the characteristics of newly emerged genes. Here we review recent genome-wide DNA and EST analyses, and discuss the three major molecular mechanisms of gene formation: (1) atypical spicing, both within and between genes, followed by adaptation, (2) tandem and interspersed segmental duplications, and (3) retrotransposition events.

A single adeno-associated virus (AAV)-murine factor VIII vector partially corrects the hemophilia A phenotype.

A major obstacle for delivery of factor (F)VIII using adeno-associated virus (AAV) vectors is the large size of FVIII cDNA, which is well above the 5 kb packaging limit for AAV. Here we construct a < 5 kb FVIII-AAV vector using murine FVIII cDNA and a strong liver-specific albumin promoter. We assessed the efficacy of this vector using three different routes of administration, intraportal, intrasplenic and tail vein injection, in FVIII knockout (FVIII KO) mice. The peak level of FVIII observed was about 8% of normal mouse FVIII activity. Even at 9 months, post vector injection, 14 of 19 mice receiving FVIII-AAV demonstrated phenotypic correction and roughly 2% FVIII activity. The transgene copy number ranged from 0.001 to 0.1 copies per cell, depending upon the somatic tissue. The potential for germline transmission of AAV was assayed in 34 pups obtained from five pairs of treated, phenotypically corrected adult hemophilic mice. Although the parents harbored the transgene in liver, spleen, and gonads, none of the 34 offspring was positive for the transgene, suggesting that the risk of inadvertent germline transmission is low.

Twin priming: a proposed mechanism for the creation of inversions in L1 retrotransposition.

L1 retrotransposons are pervasive in the human genome. Approximately 25% of recent L1 insertions in the genome are inverted and truncated at the 5' end of the element, but the mechanism of L1 inversion has been a complete mystery. We analyzed recent L1 inversions from the genomic database and discovered several findings that suggested a mechanism for the creation of L1 inversions, which we call twin priming. Twin priming is a consequence of target primed reverse transcription (TPRT), a coupled reverse transcription/integration reaction that L1 elements are thought to use during their retrotransposition. In TPRT, the L1 endonuclease cleaves DNA at its target site to produce a double-strand break with two single-strand overhangs. During twin priming, one of the overhangs anneals to the poly(A) tail of the L1 RNA, and the other overhang anneals internally on the RNA. The overhangs then serve as primers for reverse transcription. The data further indicate that a process identical to microhomology-driven single-strand annealing resolves L1 inversion intermediates.

Biology of mammalian L1 retrotransposons.

L1 retrotransposons comprise 17% of the human genome. Although most L1s are inactive, some elements remain capable of retrotransposition. L1 elements have a long evolutionary history dating to the beginnings of eukaryotic existence. Although many aspects of their retrotransposition mechanism remain poorly understood, they likely integrate into genomic DNA by a process called target primed reverse transcription. L1s have shaped mammalian genomes through a number of mechanisms. First, they have greatly expanded the genome both by their own retrotransposition and by providing the machinery necessary for the retrotransposition of other mobile elements, such as Alus. Second, they have shuffled non-L1 sequence throughout the genome by a process termed transduction. Third, they have affected gene expression by a number of mechanisms. For instance, they occasionally insert into genes and cause disease both in humans and in mice. L1 elements have proven useful as phylogenetic markers and may find other practical applications in gene discovery following insertional mutagenesis in mice and in the delivery of therapeutic genes.

A novel active L1 retrotransposon subfamily in the mouse.

Unlike human L1 retrotransposons, the 5' UTR of mouse L1 elements contains tandem repeats of approximately 200 bp in length called monomers. Multiple L1 subfamilies exist in the mouse which are distinguished by their monomer sequences. We previously described a young subfamily, called the T(F) subfamily, which contains approximately 1800 active elements among its 3000 full-length members. Here we characterize a novel subfamily of mouse L1 elements, G(F), which has unique monomer sequence and unusual patterns of monomer organization. A majority of these G(F) elements also have a unique length polymorphism in ORF1. Polymorphism analysis of G(F) elements in various mouse subspecies and laboratory strains revealed that, like T(F), the G(F) subfamily is young and expanding. About 1500 full-length G(F) elements exist in the diploid mouse genome and, based on the results of a cell culture assay, approximately 400 G(F) elements are potentially capable of retrotransposition. We also tested 14 A-type subfamily elements in the assay and estimate that about 900 active A elements may be present in the mouse genome. Thus, it is now known that there are three large active subfamilies of mouse L1s; T(F), A, and G(F), and that in total approximately 3000 full-length elements are potentially capable of active retrotransposition. This number is in great excess to the number of L1 elements thought to be active in the human genome.

Stable integration of transgenes delivered by a retrotransposon-adenovirus hybrid vector.

Helper-dependent adenoviruses show great promise as gene delivery vectors. However, because they do not integrate into the host chromosome, transgene expression cannot be maintained indefinitely. To overcome these limitations, we have inserted an L1 retrotransposon/transgene element into a helper-dependent adenovirus to create a novel chimeric gene delivery vector. Efficient adenovirus-mediated delivery of the L1 element into cultured human cells results in subsequent retrotransposition and stable integration of the transgene. L1 retrotransposition frequency was found to correlate with increasing multiplicity of infection by the chimeric vector, and further retrotransposition from newly integrated elements was not observed on prolonged culture. Therefore, this vector, which utilizes components of low immunogenic potential, represents a novel two-stage gene delivery system capable of achieving high titers via the initial helper-dependent adenovirus stage and permanent transgene integration via the retrotransposition stage.

Subfamilies of CR1 non-LTR retrotransposons have different 5'UTR sequences but are otherwise conserved.

CR1 elements and CR1-related (CR1-like) elements are a novel family of non-LTR retrotransposons that are found in all vertebrates (reptilia, amphibia, fish, and mammals), whereas more distantly related elements are found in several invertebrate species. CR1 elements have several features that distinguish them from other non-LTR retrotransposons. Most notably, their 3' termini lack a polyadenylic acid (poly A) tail and instead contain 2-4 copies of a unique 8 bp repeat. CR1 elements are present at approximately 100,000 copies in the chicken genome. The vast majority of these elements are severely 5' truncated and mutated; however, six subfamilies (CR1-A through CR1-F) are resolved by sequence comparisons. One of these subfamilies (i.e. CR1-B) previously was analyzed in detail. In the present study, we identified several full-length elements from the CR1-F subfamily. Although regions within the open reading frames and 3' untranslated regions of CR1-F and CR1-B elements are well conserved, their respective 5' untranslated regions are unrelated. Thus, our results suggest that new CR1 subfamilies form when elements with intact open reading frames acquire new 5' UTRs, which could, in principle, function as promoters.

Human L1 retrotransposition: cis preference versus trans complementation.

Long interspersed nuclear elements (LINEs or L1s) comprise approximately 17% of human DNA; however, only about 60 of the approximately 400,000 L1s are mobile. Using a retrotransposition assay in cultured human cells, we demonstrate that L1-encoded proteins predominantly mobilize the RNA that encodes them. At much lower levels, L1-encoded proteins can act in trans to promote retrotransposition of mutant L1s and other cellular mRNAs, creating processed pseudogenes. Mutant L1 RNAs are mobilized at 0.2 to 0.9% of the retrotransposition frequency of wild-type L1s, whereas cellular RNAs are mobilized at much lower frequencies (ca. 0.01 to 0.05% of wild-type levels). Thus, we conclude that L1-encoded proteins demonstrate a profound cis preference for their encoding RNA. This mechanism could enable L1 to remain retrotransposition competent in the presence of the overwhelming number of nonfunctional L1s present in human DNA.

Genetic heterogeneity in schizophrenia: stratification of genome scan data using co-segregating related phenotypes.

Despite considerable effort to identify susceptibility loci for schizophrenia, none have been localized. Multiple genome scans and collaborative efforts have shown evidence for linkage to regions on chromosomes 1q, 5q, 6q, 8p, 13q, 10p and 22q.(1-9) Heterogeneity is likely. We previously mapped schizophrenia susceptibility loci (SSL) to chromosomes 13q32 (P = 0.00002) and 8p21-22 (P= 0.0001) using 54 multiplex pedigrees and suggested linkage heterogeneity. We have now stratified these families based on co-segregating phenotypes in non-schizophrenic first degree relatives (schizophrenia spectrum personality disorders (SSPD); psychotic affective disorders (PAD)). Genome scans were conducted for these phenotypic subgroups of families and broadened affected phenotypes were tested. The SSPD group provided its strongest genome-wide linkage support for the chromosome 8p21 region (D8S1771) using either narrow (non-parametric lod (NPL) P= 0.000002) or broadened phenotypes (NPL P = 0.0000008) and a new region of interest on 1p was identified (P = 0.006). For PAD families, the peak NPL in the genome scan occurred on chromosome 3p26-p24 (P = 0.008). The identification of multiple susceptibility loci for schizophrenia may be enhanced by stratification of families using psychiatric diagnoses of the non-schizophrenic relatives.

Genetics. L1 retrotransposons shape the mammalian genome.

What are all of those retrotransposons doing buzzing about in our genome? According to Kazazian, in his Perspective this week, these mobile pieces of DNA are busy reshaping our genome, making it more diverse and enabling us to survive and thrive through the vagaries of evolution. And just how do they do this?...zip to page 1152 to find out.

Correlation between factor VIII genotype and inhibitor development in hemophilia A.

The development of neutralizing antibodies, or inhibitors, against infused factor VIII represents a significant complication of treatment for hemophilia A. Although it is likely that both genetic and environmental factors influence whether patients form inhibitors, correlations between types of factor VIII mutations and inhibitor development are becoming apparent. Approximately 20% of all patients with severe hemophilia A generate inhibitors. Of these inhibitor patients, 90% have inversions, large deletions or nonsense mutations of the factor VIII gene that would be predicted to eliminate production of factor VIII antigen. In contrast to patients with severe disease, inhibitor formation in patients with mild/moderate hemophilia A is rare. Inhibitor patients with mild/moderate disease typically have missense mutations that may cause local conformational changes within immunogenic domains of factor VIII and lead to production of dysfunctional antigen. Taken together, hemophilia A patients are predisposed to inhibitor development with mutations causing infused factor VIII to be perceived as either 1) a completely novel antigen or 2) an immunologically altered antigen.

Partial correction of murine hemophilia A with neo-antigenic murine factor VIII.

We have previously reported a factor VIII knockout (FVIII KO) mouse model for hemophilia A. Here we demonstrate the presence of nonfunctional heavy chain factor VIII protein in the mouse, making it an excellent model for cross-reacting material (CRM)-positive hemophilia A patients, who express normal levels of a dysfunctional FVIII protein. We attempted to correct these mice phenotypically by transduction of wild-type mouse factor VIII cDNA delivered in an E1/E3-deleted adenoviral vector by tail vein injection. All treated mice displayed initial high-level FVIII expression that diminished after 1 month. Ten of 12 mice administered between 6 x 10(9) and 1 x 10(11) particles/mouse along with anti-CD4 antibody showed long-term FVIII activity (0.03-0.05 IU/ml, equivalent to 3-5% of normal FVIII) that corrected the phenotype. Wild-type murine FVIII was a neo-antigen to the KO mice, generating both cytotoxic and humoral immune responses. Immune suppression with anti-CD4 antibody abrogated these immune responses. These data demonstrate that despite the presence of endogenous FVIII protein the immune system still recognizes a species-specific transgene protein as a neo-antigen, eliciting a cytotoxic T cell response. This phenomenon may exist in the treatment of other genetic disorders by gene therapy.

Correction of the coagulation defect in hemophilia A mice through factor VIII expression in skin.

To test the hypothesis that factor VIII expressed in the epidermis can correct hemophilia A, we generated transgenic mice in a factor VIII-deficient background that express human factor VIII under control of the involucrin promoter. Mice from 5 transgenic lines had both phenotypic correction and plasma factor VIII activity. In addition to the skin, however, some factor VIII expression was detected in other tissues that have stratified squamous epithelia. To determine whether an exclusively cutaneous source of factor VIII could correct factor VIII deficiency, we grafted skin explants from transgenic mice onto mice that are double knockouts for the factor VIII and RAG-1 genes. Two graft recipients had plasma factor VIII activity of 4% to 20% of normal and improved whole blood clotting compared with factor VIII-deficient mice. Thus, expression of factor VIII from the epidermis can correct hemophilia A mice, thereby supporting the feasibility of cutaneous gene therapy for systemic disease. (Blood. 2000;95:2799-2805)

Transduction of 3'-flanking sequences is common in L1 retrotransposition.

Active LINE-1 (L1) elements possess the ability to transduce non-L1 DNA flanking their 3' ends to new genomic locations. Occasionally, the 3' end processing machinery may bypass the L1 polyadenylation signal and instead utilize a second downstream polyadenylation site. To determine the frequency of L1-mediated transduction in the human genome, we selected 66 previously uncharacterized L1 sequences from the GenBank database. Fifteen (23%) of these L1s had transposed flanking DNA with an average transduction length of 207 nucleotides. Since there are approximately 400 000 L1 elements, we estimate that insertion of transduced sequences alone may have enlarged the diploid human genome as much as 19 Mb or 0.6%. We also examined 24 full-length mouse L1s and found two long transduced sequences. Thus, L1 retrotransposition in vivo commonly transduces sequence flanking the 3' end of the element.

Determination of L1 retrotransposition kinetics in cultured cells.

L1 retrotransposons are autonomous retroelements that are active in the human and mouse genomes. Previously, we developed a cultured cell assay that uses a neomycin phosphotransferase ( neo ) retrotransposition cassette to determine relative retrotransposition frequencies among various L1 elements. Here, we describe a new retrotransposition assay that uses an enhanced green fluorescent protein (EGFP) retrotransposition cassette to determine retrotransposition kinetics in cultured cells. We show that retrotransposition is not detected in cultured cells during the first 48 h post-transfection, but then proceeds at a continuous high rate for at least 16 days. We also determine the relative retrotransposition rates of two similar human L1 retrotransposons, L1(RP)and L1.3. L1(RP)retrotransposed in the EGFP assay at a rate of approximately 0.5% of transfected cells/day, approximately 3-fold higher than the rate measured for L1.3. We conclude that the new assay detects near real time retrotransposition in a single cell and is sufficiently sensitive to differentiate retrotransposition rates among similar L1 elements. The EGFP assay exhibits improved speed and accuracy compared to the previous assay when used to determine relative retrotransposition frequencies. Furthermore, the EGFP cassette has an expanded range of experimental applications.

Mobile elements and the human genome.

Genomic DNA is often thought of as the stable template of heredity, largely dormant and unchanging, apart from perhaps the occasional point mutation. But it has become increasingly clear that DNA is dynamic rather than static, being subjected to rearrangements, insertions and deletions. Much of this plasticity can be attributed to transposable elements and their genomic relatives.

Full-length human L1 insertions retain the capacity for high frequency retrotransposition in cultured cells.

Functional L1 elements are autonomous retrotransposons that can insert into human genes and cause disease. To date, 10 of 12 known L1 retrotranspositions into human genes have been found to be 5"-truncated and incapable of further retrotransposition. Here we report the nucleotide sequences of the two full-length L1 elements, L1beta-thal and L1RP, that have inserted into the beta-globin and retinitis pigmentosa-2 (RP2) genes, respectively. L1beta-thal is 99. 4% identical to a consensus sequence of active human L1s, while L1RP is 99.9% identical. Both elements retain impressive capacity for high frequency retrotransposition in cultured HeLa cells. Indeed, L1RP is the most active L1 isolated to date. Our data indicate that not all L1 insertions into human genes are 'dead on arrival'. Our findings also lend further credence to the concept of cis preference, that the proteins encoded by a particular L1 preferentially act upon their encoding RNA as opposed to other L1 RNAs.

Lack of linkage or association between schizophrenia and the polymorphic trinucleotide repeat within the KCNN3 gene on chromosome 1q21.

To determine the importance of a candidate gene KCNN3 (formerly named hSKCa3) in the susceptibility to schizophrenia, we have studied the genotypes of a (CAG)n polymorphism within this gene in the DNAs of the members of 54 multiplex families with this disease. Parametric and nonparametric linkage analysis did not provide evidence for linkage between KCNN3 (that we mapped to chromosome 1q21) and schizophrenia. Furthermore, we observed no difference in the distribution of the (CAG)n alleles between affected and normal individuals. These results do not support the hypothesis that larger KCNN3 alleles are preferentially associated with schizophrenia [Chandy et al. 1998 Mol Psychiatr 3:32-37] in individuals from multiply affected families.