The leader intron of AtMHX can elicit, in the absence of splicing, low-level intron-mediated enhancement that depends on the internal intron sequence

Abstract

Background: Introns stimulate gene expression in a wide range of organisms by increasing the levels of mature mRNA, without affecting mRNA stability. Although introns sometimes function as transcriptional enhancers, they usually stimulate expression by a process termed intron-mediated enhancement (IME). The mechanism of IME is largely unknown. While splicing per se is not sufficient for IME, as evident from the fact that not all introns increase expression, it is not clear yet whether splicing of the enhancing introns is essential for enhancement. The leader intron (LI) of the Arabidopsis AtMHX gene was previously shown to substantially increase the expression of the AtMHX promoter. Here we investigated whether this LI acts as a transcriptional enhancer and whether its splicing is essential for IME.

Results: Expression in transformed Arabidopsis plants of an AtMHX::GUS construct from which the LI was eliminated was similar to a construct that included only the minimal promoter fused to GUS. Yet, almost no expression was seen in constructs that included the LI in addition to the minimal promoter or the LI inserted in various locations in the promoter. While the LI enhanced 272-fold the expression of the weak AtMHX promoter, only a 3-fold enhancement was observed for the strong CaMV 35S promoter. In the context of the AtMHX promoter, an unspliceable version of the LI that had mutated 5' and 3' splice sites mediated a low-level (5-fold) enhancement. Eliminating the internal 320 nt of the 416 nt unspliceable intron resulted in loss of ability to mediate low-level enhancement.

Conclusions: Although AtMHX promoter shows almost no expression in the absence of its LI, this intron does not act as a transcriptional enhancer and is unable to support expression in the absence of the enhancer elements of the promoter. It is also shown that the same intron can have very different contributions to expression of different promoters. Our results also demonstrate that while splicing is essential for substantial IME, in the absence of splicing low-level enhancement can be obtained. Notably, it is shown that the internal intron sequence plays a significant role in mediating the low-level enhancement of unspliced introns.

Figure 1

Schematic representation of the constructs utilized. The AtMHX promoter elements utilized were the repetitive element (RE), unique sequence (US), and minimal promoter (MP) (see Results and Methods for details about these elements). The solid line represents the 5'UTR of AtMHX, the diagonal lines represent the leader intron (LI) of AtMHX, and the dashed line represents the region eliminated from the LI. GUS - the coding sequence of β-glucuronidase; AtMHX ter - the terminator of the AtMHX gene; 35S - the CaMV 35S promoter; NOS ter - the terminator of the Agrobacterium tumefaciens nopaline synthase gene; X - an abolished splice site. The objective of each construct set is explained in the text.

Figure 2

The LI does not act as a transcriptional enhancer. A. Mean and standard errors of relative GUS activity in mixtures including an equal number of T2 seedlings from 13 independent transformants of each construct. GUS activity of the WT construct was assigned the value of 1. EV, plants expressing the empty vector. The internal graph shows relative GUS activity compared to the WT construct on a smaller scale in plants expressing the constructs that mediated low GUS activity. Statistical analysis (t-test) revealed a highly significant difference (p < 0.01) between plants expressing the WT and all other constructs, but no significant difference was revealed between plants expressing the -Int and the MP+I, MP, I+MP, PIa, or PIb constructs (p > 0.05). B. Representative Northern blot results with the GUS probe. rRNA, ribosomal RNA.

Figure 3

A 3-fold increase in expression level is seen in the context of the 35S promoter. A. Mean and standard errors of relative GUS activity in six biological replicates, each including an equal number of T2 seedlings from 20 independent transformants of each construct. GUS activity of the 35S construct was assigned the value of 1. B. Representative Northern blot results with the GUS probe. rRNA, ribosomal RNA. C. Mean and standard errors of relative GUS transcript levels in the six biological replicates. Quantification of band densities on gel was performed with the ImageJ program. The data are presented relative to the transcript level of the 35S construct.

Figure 4

The modifications introduced in the different LIs. Boxed, partial sequence of the 5'UTR exons that border the LIs. Exon sequence was not altered in any of the constructs. The modifications introduced into the Wm and Wm-S constructs are highlighted in gray. The underlying numbers indicate the purpose of each modification: 1 - eliminating the main and potential cryptic splice sites, 2 - maintaining the length, deduced amino-acid sequence and predicted secondary structure of the natural uORF found in the 5'UTR of AtMHX [the first (ATG) and last (TGA) codons of the uORF of each construct are bold and underlined], 3 - creating SacII sites for obtaining the WmD-S from the Wm-S construct, 4 - eliminating the internal ATG codons, and 5 - eliminating a HindIII site that interfered with cloning into the binary vector.

Figure 5

Determination of splicing efficiency. A. The primers used for RT-PCR are indicated on the illustration of the WT construct. The forward and reverse primers were derived from the first exon of the 5'UTR and from the GUS coding sequence, respectively. B. The results of the RT-PCR analysis. The analysis was not quantitative but qualitative (that is, the analysis shows the size of each transcript, but band intensity do not accurately reflect the relative abundance of the different transcripts). All the templates were cDNA except pWT, which was the plasmid that included the WT construct. The product of pWT thus indicated the size of unspliced WT transcript. The boxes separated by a long or short line indicate the expected size of unspliced transcripts including either the full or deleted LI, respectively. The two adjacent boxes indicate the size of correctly spliced transcripts.

Figure 6

The unspliced LI mediates low-level enhancement. A. Mean and standard errors of relative GUS activity in mixtures including an equal number of T2 seedlings from 13 independent transformants of each construct. GUS activity of the Wm construct was assigned the value of 1. EV, plants expressing the empty vector. The internal graph shows relative GUS activity compared to the Wm construct on a smaller scale in plants expressing the constructs that mediated low GUS activity. Statistical analysis (t-test) revealed a highly significant difference (p < 0.01) between plants expressing the Wm and all other constructs. A highly significant difference (p < 0.01) was also observed between plants expressing the -Int and the Wm-S constructs, but not between plants expressing the -Int and the WmD-S construct. B. Representative Northern blot results with the GUS probe. rRNA, ribosomal RNA.