Thomas R. Broker

Complex splicing patterns of RNAs from the early regions of adenovirus-2

Abstract The RNA species encoded by the five early regions of adenovirus serotype 2 (Ad2) were isolated from the cytoplasm of HeLa cells at various times after infection either in the absence or in the presence of cycloheximide or cytosine arabinoside. The transcripts were characterized by electron microscopy of heteroduplexes formed with single-stranded Ad2 DNA. Each early region gave rise to a family of composite RNA species with up to four conserved segments spliced together after the deletion of intervening sequences. Transcripts from each family contained some sequences in common but differed in the lengths and positions of the conserved segments and thus had alternative splicing patterns. They were mapped on the adenovirus genome and the relative abundance of each species was determined. The map co-ordinates of the early region promoters were refined, based upon the indication that the 5′ ends of mature cytoplasmic RNAs are derived from promoter-proximal DNA sequences. The 5′ leader segment of the region 2 message for the single-stranded DNA binding protein changed during the course of infection, suggesting the utilization of different promoters at early and late times. The transition was prevented by blocking protein synthesis with cycloheximide. Transcription of region 3 was under the control of its own promoter at early times but could also be directed by the major r -strand late promoter at intermediate to late times. The internal splicing patterns of the region 3 transcripts seen at early times persisted at late times, but some molecules had the 5′ tripartite leaders common to late r -strand mRNAs. One of the major early region 3 transcripts contained three RNA segments which are also present in a presumptive processing intermediate for the fiber RNA, a late message which is separated from its promoter by early region 3. These extra segments in some fiber RNAs apparently reflect the recognition of early RNA splicing signals at late times. From the large, complex arrays of composite RNA structures, numerous insights into the RNA splicing mechanisms were inferred.

A map of cytoplasmic RNA transcripts from lytic adenovirus type 2, determined by electron microscopy of RNA:DNA hybrids

Abstract We have used electron microscopic RNA loop mapping to determine, to within 200 nucleotides, the chromosome coordinates of the 5′ and 3′ ends of adenovirus type 2 transcripts isolated from the cytoplasm of productively infected human KB cells. The major blocks of transcription and many of the individual transcripts have discrete lengths and well defined 5′ and 3′ termini. Early cytoplasmic RNA is transcribed from four separate regions of the adenovirus genome, in map intervals 1.3–11.1, 62.4–67.9, 78.6–86.1 and 91.5–96.8. The most frequent late RNA loop occurred between coordinates 51.9 and 62.2. A late transcript complementary to interval 86.3–91.5 formed a characteristic R loop with a long branch of displaced RNA from coordinate 89.8–91.5, possibly a result of secondary structure near coordinate 89.8. Other late cytoplasmic RNA loops were located between coordinates 11.1 and 14.9, and in the intervals from 30–52 and 68–83. Transcripts from several regions of the chromosome had variable lengths but common 5′ or 3′ termini. Few R loops were seen between map coordinates 15 and 30. Switches between template strands were observed at map positions 11.2, 62.4 and 91.5. Many of the R loops have been correlated with gene products mapped previously by translation of RNA complementary to different portions of the chromosome.

Underwound loops in self-renatured DNA can be diagnostic of inverted duplications and translocated sequences

Abstract DNA that contains inverted duplications separated by non-inverted sequences often can form characteristic “underwound loops” when it is denatured and reannealed. An underwound loop is a partially double-stranded, partially denatured segment between the inverted duplications and is produced as follows. During the early stages of the reannealing, intrastrand stem-loop structures form with first-order kinetics when the inverted duplications pair. In a slower second-order reaction, complementary strands (each with a stem-loop) reanneal. The stem-loop structures produce a cruciform in the hybrid. Because of the unpaired sequences in the loop, the cruciform is unstable. It can isomerize to a linear duplex by double-strand exchange of complementary sequences in the stems. This process requires co-ordinated axial rotation of the stems and the flanking duplexes as well as rotation of the loops. If, however, complementary sequences in the loops start to pair, axial rotation is prevented and the stem-loop structures are trapped in a metastable state. The strands of separate, closed rings cannot interwind when they pair. Consequently, the loops observed by electron microscopy have variable patterns of single-stranded denaturation bubbles and duplex segments with both right-handed and left-handed winding. We have used underwound loops to identify a short inverted duplication flanking the γδ recombination sequence of Escherichia coli F factor (isolated on φ80 d 3 ilv + transducing phage) and to study DNA from phages Mu and P1 in which the G segments are flanked by inverted duplications. When deproteinized adenovirus-2 DNA was denatured and reannealed, some underwound circles the length of the entire chromosome were observed by electron microscopy. These resulted from the restricted interaction of complementary single-stranded rings generated when pairing of the short inverted terminal duplications closed the ends of single strands. Another type of underwound loop was seen in heteroduplexes containing complementary insertion loops located at different positions in the hybridized strands, such as occurs with P1 cam DNAs. All these underwound structures are similar in appearance to the hybrids formed when topologically separate, complementary single-stranded circles of Colicin E 1 DNA were allowed to anneal.

Adenovirus-2 messengers--an example of baroque molecular architecture

Adenovirus type 2 (Ad2) causes respiratory infections in humans, grows productively on human cell lines such as HeLa and KB, and can transform rat primary cell lines (Tooze 1973). The Ad2 virion consists of a linear duplex DNA chromosome 35,000 base pairs (bp) long (23 × 106 daltons) contained within an icosahedral capsid composed of at least ten different proteins. After penetration of the cell membrane, the viral DNA and several core proteins associated with it are transported to the nucleus, where early RNA transcription begins within 2 hours. About 8 hours after infection, viral DNA replication commences, reaching a maximum rate several hours later. Late transcription to produce messenger RNA (mRNA) for capsid and other virus-specific proteins begins after the onset of DNA replication and continues for about 2 days, when the infected cells die.

RNA transcription and splicing at early and intermediate times after adenovirus-2 infection

Transcription of human adenovirus-2 (Ad2) DNA1 has generally been recognized to have an early phase (from 1 to 8 hours after infection of permissive human cells) and a late phase, essentially coincident with DNA replication (from 8 hours until cell death at 48–72 hours after infection). Solution-hybridization studies have shown that about one fourth of the genome is expressed as cytoplasmic RNA at early times after infection and that these RNAs come from four widely separated regions, two on each DNA strand (Green et al. 1971; Sharp et al. 1975; Pettersson et al. 1976). Electron microscopic studies of R loops formed with early Ad2 RNA in double-stranded Ad2 DNA (Chow et al. 1977b; Westphal and Lai 1977) as well as nuclease-S1 mapping of RNA-single-stranded-DNA heteroduplexes (Berk and Sharp 1977b) produced evidence that early region 1 is subdivided into two segments, 1A and 1B.

Adenovirus DNA replication in soluble extracts of infected cell nuclei.

Synthesis of viral DNA after infection of prokaryotic and eukaryotic cells has been studied to understand the replication of the viral genome itself, as well as to serve as a substrate for the analysis of components necessary for host-chromosomal replication (Kaufmann et al. 1977; Kornberg; Sumida-Yasumoto et al.; Hillenbrand et al.; Liu et al.; all this volume). Viral systems have been essential because the physical and enzymatic studies used to characterize the replication reaction are considerably more difficult on the large host DNA molecules. These models have been particularly useful in the study of the molecular mechanisms of prokaryotic DNA replication because many of the smaller phages require host-coded gene products for replication (Wickner and Hurwitz, 1975; Bouche et al. 1975; Sumida-Yasumoto et al.; Meyer et al.; both this volume).