Modern thoughts on an ancyent marinere: function, evolution, regulation.

INTRODUCTION

This is a summary review of the mariner family of transposable elements: What do we know? What do we not know that we should? Why is it important that we know it? The end of the matter is that we know the most about mechanism (much of it by inference from related transposable elements). We also know quite a lot about general features of population dynamics and evolution, although we are short on the details and lack a comprehensive model. The major gap is that we know almost nothing about regulation. That the study of regulatory mechanisms has not been pursued more aggressively is quite paradoxical, because regulation is perhaps the key issue in the application of mariner-like elements to methods of germline transformation of insect pests and vectors of human disease. Happily, experiments carried out for other purposes have recently identified several candidate mechanisms of regulation that may, singly or in combination, mediate regulation of this important family of transposable elements.

RELATION TO OTHER D,D(35)E ELEMENTS

Members of the mariner family of transposable elements (MLEs), as well as members of its presumed sister group, the Tc1 family (TLEs), are of great current interest for a number of reasons. First, they are eukaryotic members of a larger superfamily of transposable elements that includes such prokaryotic members as the bacteriophage Mu, the transposon Tn7, and many bacterial insertion sequences including the Escherichia coli elements IS2, IS3, IS4, and IS30 (19). The mariner/Tc1/IS superfamily of transposable elements is related to a still larger assemblage of sequences that includes human immunodeficiency virus (HIV) and the copia and gypsy families of long-terminal-repeat (LTR) retrotransposons (15).

What these sequences have in common is that their transposase or integrase proteins include a sequence motif called the D,D(35)E motif, which consists of two aspartic acid residues, typically separated by more than 90 amino acids, followed by a glutamic acid residue, typically 34 or 35 amino acids further toward the carboxyl end (19). Given its rather vague specification, the D,D(35)E motif is more a sort of signature than a "motif" in the usual sense of the word. Nevertheless, the signature is found in a very diverse set of proteins.

The possible evolutionary relationships between various members of the D,D(35)E superfamily have been examined by molecular phylogenetic methods (15). Even though there are not a large number of phylogenetically informative sites in the blocks of amino acids surrounding the key acidic residues of the D,D(35)E signature, the analysis does lead to three seemingly robust conclusions. First, the mariner/Tc1 superfamily, on the one hand, and the LTR retrotransposons and retroviruses, on the other hand, are separated into two monophyletic groups. Second, the bacterial IS elements in the superfamily are located willy nilly in the tree and not in a coherent group; it is not clear whether the dispersion of the IS elements results from an inadequate number of phylogenetically informative sites or whether the IS elements are truly polyphyletic. Third, there is relatively strong bootstrap support (though still less than 80%) that groups mariner elements with Tc1 elements.

The apparent common ancestry of MLEs and TLEs is all the more interesting in light of the fact that mariner is the only member of the extended D,D(35)E superfamily that does not have the D,D(35)E signature. Bear in mind that the D,D(35)E signature is not a consensus sequence in the usual sense of "majority rule" over a set of aligned sequences. Indeed, the three acidic residues are absolutely invariant across members of the Tc1 family thought to be functional (65), and the last two are invariant across members of the extended superfamily, prokaryotic IS elements included (19). In contrast, the mariner-like elements have the signature D,D(34)D (65). Does it matter? Who would bet it does? After all, glutamic acid and aspartic acid are both acidic residues, the pK values for the side-chain carboxyl groups are almost the same (4.3 versus 3.9), a change from one residue into the other is "conservative," and the amino acids are often found at corresponding sites in homologous proteins. The answer, however, is that it does matter. In mariner, the "conservative" change from D,D(34)D to D,D(34)E completely obliterates transposase activity (40).

MECHANISM OF TRANSPOSITION AND EXCISION

Much of the work on the extended D,D(35)E superfamily concerns molecular mechanisms. What is the molecular structure of the transposase? What are its DNA-binding sites and what sequences are recognized in the transposable element that is mobilized? What is the molecular mechanism of transposition? Is the transposase alone sufficient to carry out the transposition reaction?

The D,D(35)E motif is thought to be a key player in the reaction mechanism, part of the active site that serves as a binding domain for a divalent cation ([Mg.sup.2+] or [Mn.sup.2+]) necessary for catalysis (34). The unifying mechanistic feature of proteins that share this motif is the ability to execute a single-strand scission in a duplex DNA molecule that exposes a reactive 3[prime] hydroxyl (18). A single-strand scission at each end of a mobile DNA sequence is the essential D,D(35)E contribution to the simple insertion of a retrovirus or to cointegrate formation of bacteriophage Mu (55). In each case, the reactive 3' hydroxyl groups are joined with nucleotides at displaced positions on opposite strands of a target sequence, which, when repaired by host enzymes, creates the direct duplication characteristic of transposable element insertions.

The transposase reactions of many Class II transposable elements initiate with a staggered cleavage of both strands at each end of the transposable element. The entire element is thereby released from the donor molecule and free to insert by a "cut-and-paste" mechanism into a target site, again through initial joining of the 3[prime] overhanging ends of the transposable element (18). The 5[prime] overhang created at the vacated target site apparently forms a short heteroduplex, which is repaired by the host mismatch repair system and results in a characteristic "footprint" after excision. First established for Tn7 (17) and Tn10 (33), the cut-and-paste mechanism is also the mode of transposition of Tc1-related elements. For Tc3 and for one mariner-like element, the cut-and-paste mechanism has been demonstrated in vitro (35, 76, 77).

Based on general similarity with Tc1-related elements, the presumption is very strong that all mariner elements also transpose by a cut-and-paste mechanism. The strongest indirect evidence comes from the footprints left after mariner excision. The experimental assay is based on a two-component experimental system in Drosophila. One component consists of the Mos1 (mosaic-1) mariner element, which supplies the functional transposase. This element is a full-length element of 1286 base pairs (hp), including the 28-bp inverted repeats, containing an uninterrupted open reading frame that codes for a putative transposase of 345 amino acids (52, 54). The functional organization of the Mos1 element is illustrated in Figure 1. The locations of the D,D(34)D signature residues are indicated. A bipartite nuclear localization signal has also been identified by sequence similarity to a bona fide nuclear localization signal present in certain TLEs in zebrafish (26). The middle R of the RRK region is conserved in both MLEs and TLEs (65); genetic evidence that it is essential comes from an R-to-H mutation in the Mos1 transposase, which completely eliminates activity (40).

The second component of the transposase assay is an inactive mariner element, called peach, which is also a full-length element but one coding for an inactive transposase (22, 27). The peach element differs from Mos1 at 11 nucleotide sites, including four amino acid replacements (52, 53). Presence of the peach element is detected through its phenotypic effects when inserted in the 5[prime] noncoding region of the X-linked white gene; the peach insertion in the wpch transgene results in peach-colored eyes. Excision of the peach element from wpch by functional transposase supplied in trans restores wild-type expression of the white gene. In the soma, excision results in eye-color mosaicism (red spots on a peach background); in the germline, it results in a reverse mutation of wpch to wild type (9, 10).

In the wpch excision assay, the Mos1 element is highly active, resulting in eye-color mosaicism in all flies that carry it (52, 54). The type of footprints remaining after peach excision are exactly analogous to those …