Post-entrapment genome engineering: first exon size does not affect the expression of fusion transcripts generated by gene entrapment

Abstract

Gene trap mutagenesis in mouse embryonic stem cells has been widely used for genome-wide studies of mammalian gene function. However, while large numbers of genes can be disrupted, individual mutations may suffer from limitations due to the structure and/or placement of targeting vector. To extend the utility of gene trap mutagenesis, replaceable 3' [or poly(A)] gene trap vectors were developed that permit sequences inserted in individual entrapment clones to be engineered by Cre-mediated recombination. 3' traps incorporating different drug resistance genes could be readily exchanged, simply by selecting for the drug-resistance gene of the replacement vector. By substituting different 3' traps, we show that otherwise identical fusion genes containing a large first exon (804 nt) are not expressed at appreciably lower levels than genes expressing small first exons (384 and 151 nt). Thus, size appears to have less effect on the expression and processing of first exons than has been reported for internal exons. Finally, a retroviral poly(A) trap (consisting of a RNA polymerase II promoter, a neomycin-resistance gene, and 5'-splice site) typically produced mutagenized clones in which vector sequences spliced to the 3'-terminal exons of cellular transcription units, suggesting strong selection for fusion transcripts that evade nonsense-mediated decay. The efficient exchange of poly(A) traps should greatly extend the utility of mutant libraries generated by gene entrapment and provides new strategies to study the rules that govern the expression of exons inserted throughout the genome.

Figure 1.

Tagged sequence mutagenesis with the LNPAT1 gene trap vector. An LNPAT1 provirus is shown integrated after the second exon (black) of a hypothetical gene. The Neo-resistance gene (dark gray shading), expressed from the RNA polymerase II (PolII) promoter, splices to downstream exons (III and IV). Transcripts of the disrupted gene, containing exons I and II, splice to coding sequences for the enhanced green fluorescent protein (EGFP) preceded by the ECMV internal ribosome entry site (IRES) in cells where the occupied gene is expressed. Cellular sequences appended to the 3′-end of virus-cell fusion transcripts are amplified by 3′-RACE and sequenced. The resulting sequence tags identify genes disrupted by the provirus and are subcloned for later use, for example, to prepare DNA microarrays. The resulting library of mutant ES cell clones, cryopreserved in liquid nitrogen, provides a source of mutant alleles that may be transmitted into the mouse germ line. A Herpes virus thymidine kinase gene (TK) and RNA destabilization element (flash symbol) are also present. Heterospecific loxP sites (loxP and loxP5171) allow vector sequences to be replaced by Cre-mediated cassette exchange (see text for details).

Figure 2.

Replaceable 3′ gene entrapment cassettes. (A,B) neomycin- and (C,D) zeocin-resistance genes ending at a 3′-splice site (SD) and RNA destabilization sequence (flash symbol) were cloned downstream of the (A) RNA polymerase II (Pol II) or (B,C,D) phosphoglycerol kinase (PGK) promoters. IntZeoPTC contains a synthetic intron inserted within the Zeocin-resistance gene (D). After gene entrapment, vector sequences splice to downstream exons of cellular genes (black boxes). The sizes of 5′-exons containing Neo and Zeo sequences is indicated in base pairs (bp). Heterospecific loxP sites (loxP and lox5171) allow vector sequences to be replaced by Cre-mediated cassette exchange.

Figure 3.

Tagged-sequences mutagenesis with the LNPAT1 3′ gene trap. Here 172 vector-fusion transcripts cloned by 3′-RACE corresponded to genes disrupted in ES cells by the LNPAT1 vector as determined by their DNA sequence (confirmed 3′-RACE products). Of these, 151 provided unique sequence tags (i.e., they were not derived from sister clones from the same culture dish), and 143 matched the mouse genome sequence (MGS). Based on the mouse genome sequence annotation, 68 (48%) of the RACE sequences matched transcribed sequences (i.e., exons), corresponding to 22 previously characterized genes, 18 Riken cDNAs, 10 transcription units identified by gene prediction software (NCBI LOC genes), and 18 EST contigs not included among the Entrez gene annotations (ESTs). Of the remaining 53 sequence tags, eight matched intron sequences within annotated genes, 45 matched genomic sequences not contained within annotated transcription units and therefore may be between genes, and 22 contained repetitive sequences and could not be placed on the genome sequence.

Figure 4.

Cre-mediated cassette exchange at the 4930562A09Rik locus. (A) Sequences of the LNPAT1 provirus inserted in the 4930562A09Rik locus were replaced by the ZeoPTC gene trap cassette, which was then replaced by NeoPTC. (B) Clones containing the proper gene replacements were identified by Southern blot analysis. The replacement of LNAPT1 by ZeoPTC (upper panel) was accompanied by the loss of the 4.5-kb Neo-hybridizing SacI fragment and the presence of a 3.2-kb Zeo-hybridizing fragment (lanes 4,11). SacI sites are located in each LTR and in the Tk gene. The replacement of ZeoPTC by NeoPTC (lower panel) was accompanied by the loss of the 3.2-kb Zeo-hybridizing fragment and the presence of a 3.6-kb Neo-hybridizing fragment (clones 7, 8, 9, 13, 15, 18, 19, 21, and 23).

Figure 5.

Fusion gene expression in replacement clones. RNA from cell clones expressing poly(A) traps inserted in the (A) 4930562A09Rik,(B) LOC228098, and (C) 4933407O12Rik genes were analyzed by Northern blot hybridization, probing with gene-specific, Neo-specific, and Zeo-specific probes. The clones in A were obtained by replacing an LNPAT1 poly(A) trap (Pol2Neo) inserted in the 4930562A09Rik gene with ZeoPTC (PGKZeo), and then by replacing the ZeoPTC cassette with NeoPTC (PGKNeo). The clones in B were obtained by replacing a NeoPTC poly(A) trap in the LOC228098 gene with either ZeoPTC (PGKZeo) or IntZeoPTC (PGKIntZeo). The clones in C were obtained by replacing a IntZeoPTC (PGKIntZeo) poly(A) trap in the 4933407O12Rik gene with NeoPTC (PGKNeo). All clones expressed fusion transcripts of the expected size as indicated on the left. None of the targeted genes appeared to be expressed in the parental stem cells (ES), confirming the ability of poly(A) traps to disrupt nonexpressed genes.

Figure 6.

Stability of Neo- and Zeo-4933407O12Rik fusion transcripts. (A) Northern blot analysis of RNA decay in actinomycin D-treated cells. RNAs were extracted from actinomycin D-treated cells expressing Neo-4933407O12Rik and Zeo-4933407O12Rik fusion transcripts, and the Northern blots were hybridized to a 4933407O12Rik-specific probe. RNAs were also probed with a β-actin probe to assess relative levels of RNA in each sample. (B) Kinetics of fusion transcript turnover. Relative levels of Neo-4933407O12Rik (squares) and Zeo-4933407O12Rik (circles) fusion transcripts in A were measured by PhosphorImager densitometry and plotted as a function of time. Zeo-4933407O12Rik fusion transcripts were twice as stable as the otherwise identical Neo fusion transcripts.