Integrated functional and bioinformatics approach for the identification and experimental verification of RNA signals: application to HIV-1 INS

Abstract

Regulation of gene expression involves sequence elements in nucleic acids. In promoters, multiple sequence elements cooperate as functional modules, which in combination determine overall promoter activity. We previously developed computational tools based on this hierarchical structure for in silico promoter analysis. Here we address the functional organization of post-transcriptional control elements, using the HIV-1 genome as a model. Numerous mutagenesis studies demonstrate that expression of HIV structural proteins is restricted by inhibitory sequences (INS) in HIV mRNAs in the absence of the HIV-1 Rev protein. However, previous attempts to detect conserved sequence patterns of HIV-1 INS have failed. We defined four distinct sequence patterns for inhibitory motifs (weight matrices), which identified 22 out of the 25 known INS as well as several new candidate INS regions contained in numerous HIV-1 strains. The conservation of INS motifs within the HIV genome was not due to overall sequence conservation. The functionality of two candidate INS regions was analyzed with a new assay that measures the effect of non-coding mRNA sequences on production of red fluorescent reporter protein. Both new INS regions showed inhibitory activity in sense but not in antisense orientation. Inhibitory activity increased by combining both INS regions in the same mRNA. Inhibitory activity of known and new INS regions was overcome by co-expression of the HIV-1 Rev protein.

Figure 1

HIV-1 INS elements, INS regions, genomic organization and gene expression and weight matrix generation scheme. (a) The corresponding INS element names are depicted above the HIV genome. Reading frames [p17^gag, p24^gag, p15^gag, protease, reverse-transcriptase (p66RT), integrase, gp120^env, gp41^env and several accessory proteins] and the corresponding INS regions (INS1, INS2, INS3 and CRS) containing experimentally identified and verified INS elements are indicated above the two HIV-1 transcript classes. (b) An example of the matrix generation scheme, which is described in detail in Results.

Figure 2

Strategy for the generation of INS weight matrices (INS motifs). This figure shows a schematic representation of the general strategy for the identification of the HIV-1 INS motifs (from top to bottom). The approach started in the laboratory with the detailed mutational analysis, which is indicated in the right half of the figure while the multiple steps of the computational, bioinformatics approach are represented on the left. Arrows between both sites indicate crosscheck analyses between experimental data and the computational results. Circular iterations were necessary between steps II and III and within step III (see text for details).

Figure 3

Distinct internal composition of INS regions in the GAG/POL-ORF of HXB2R. To the right of the weight matrix name, INS motif matches are indicated as symbols as described in Figure 4. The corresponding INS element names are depicted above and the INS regions indicated below. Reading frames [p17^gag, p24^gag, p15^gag, protease, reverse-transcriptase (p66RT), integrase] and the corresponding INS regions (INS1, INS2, INS3 and CRS) containing experimentally identified and verified INS elements are indicated.

Figure 4

Comparative mapping of INS regions and motifs in 26 full-length HIV-1 proviral sequences. In this graphical representation of the alignment of 26 HIV-1 full-length proviral genome sequences, the top line illustrates the organization of the HIV-1 genome (ORFs, line 1). Line 2 shows the localization of the INS regions (rectangles), with the known regions marked in orange and the predicted regions uncolored. Line 3 shows the positions of the INS elements (orange boxes) in the HIV-1 HXB2 genome. INS elements were identified by loss of Rev-dependent expression after mutation (27). The lines below represent the following individual HIV-1 sequences (listed from top to bottom, with the corresponding clades indicated in parentheses): hxb2r (B), lw123 (B), 3202a12 (B), 3202a21 (B), mn (B), cam1 (B), th475a (B), bcsg3 (B), lai (B), weau160 (B), han (B), d31 (B), jrcsf (B), jrfl (B), manc (B), oyi (B), ny5cg (B), rf (B), ndk (D), z2z6 (D), eli (D), u455 (A), mal (ADI), ibng (AG), mvp5180 (O) and ant70 (O). Different geometrical shapes and colors indicate matches to the INS matrices (INS motifs). INS motifs for INS-A are marked as red diamonds, for INS-B as filled skyblue squares, for INS-MC as green circles, and for INS-MD as orange boxes. Since INS elements and INS regions are functional on the mRNA level only the sense strand (plus strand) is shown. Open boxes below the aligned sequences indicate the newly predicted INS regions (INS4, INS5, INS6, INS7, INS8, INS9, INS10). A complete and more detailed version of this Figure is available on our homepage (www.gsf.de/biodv/).

Figure 5

Reporter plasmids for functional INS analysis. The construct pLRedR contains the HIV 5′-LTR as promoter/enhancer, an ORF encoding a red fluorescent reporter protein (RFP; DsRed1), multiple additional translational STOP codons for termination of the DsRed ORF, the RRE and the 3′-LTR of HIV for termination of transcription. Different INS regions (Fig. 4) were inserted in sense and antisense orientation into the unique ClaI restriction site of the basic reporter plasmid pLRedR.

Figure 6

Evaluation of the influence of INS regions on expression of red fluorescent reporter protein by flow cytometry. HLtat cell cultures were transfected with different RFP reporter plasmids (see Fig. 5) and a GFP-expression plasmid as transfection control. The fluorescence of cells [shown for single cells in (a)] was evaluated by flow cytometry (b). The y-axis represents red fluorescence, reflecting expression of the reporter protein from the gene with the INS region, whereas the x-axis displays green fluorescence of the transfection control protein. Fewer cells in the upper right quadrant indicate a greater down-modulating effect of the INS region in the construct. Quantitative evaluation was carried out based on the histogram data shown in (c). Counts indicate cell numbers (y-axis) and the x-axis represents fluorescence intensity. The median of red fluorescence intensity was determined from the range labeled R2 (black bars). The gray areas were excluded as background (level observed with non-transfected cells).

Figure 7

Inhibitory effects of known and predicted INS regions. To analyze inhibitory activity of known and predicted INS regions of the HIV-1 genome, cells were transfected in parallel with various reporter plasmids containing INS regions or with the pLRedR control plasmid, which lacks INS (plasmids shown in Fig. 5). Flow cytometry was used to determine the ratio of green to red fluorescence (detailed in Materials and Methods). For each INS-bearing construct, this ratio was normalized to the value of pLRedR, which was set to 1. Inhibitory activity was evaluated for the known INS regions, INS1 and INS2, individually and in combination (INS1+2) (left side) and similarly for the newly predicted INS region candidates (INS5 and INS6) and their combination (right side). All INS regions were analyzed in sense (black bars, s) and antisense orientation (gray bars, as). FACS data were obtained from five independent experiments, all of which are shown in the table below the graph. The statistical significance of the difference between sense and antisense was determined as detailed in Material and Methods (Mann–Whitney non-parametric test) and is given below the table. The graph shows the values of experiment 1.

Figure 8

Rev rescue of INS inhibitory activities. To analyze whether the inhibitory activity of known and predicted INS regions of the HIV-1 genome can be overcome by HIV-1 Rev protein, reporter plasmids with the indicated INS regions were cotransfected in parallel with either pCsRevsg143 for expression in the presence of Rev-GFP or pFRED143 for expression in the absence of Rev. Reporter plasmid pLRedR, which lacks INS (Fig. 5), was analyzed as a control. Cells showing green and red fluorescence were counted and the Rev-induced increase in the number of dual-fluorescent cells determined (detailed in Materials and Methods). The effect of Rev on all INS regions was analyzed in sense (black bars, s) and antisense orientations (gray bars, as).