Introduction

Tissue Factor (Hugo Gene Nomenclature Consortium HGNC gene Factor III (F3)) plays vital roles in many cellular processes including haemostasis, thrombosis, inflammation and angiogenesis1. It has a varied means of expression across different cell types. In vascular adventitial fibroblasts and epithelial cells, it is expressed constitutively and may act as a haemostatic envelope triggering coagulation in the event of tissue injury2.

The intracellular trafficking of F3 allows rapid mobilisation to the cell surface when appropriate. Whilst much is known about the expression pattern of F3 on the cell surface of different cell types, the subcellular localisation and dynamics have been less studied. In both growth factor stimulated smooth muscle cells and Baby Hamster Kidney cells transfected with exogenous F3, the majority of F3 protein localises to intracellular pools in the perinuclear cytoplasm3,4. Mandal et al. showed that a large fraction of the intracellular pool of F3 in human fibroblasts is localised to the Golgi5. Furthermore, they found that activation of protease-activated receptor-2 (PAR-2), a G-protein coupled receptor involved in various cellular responses, mobilises intracellular F3 from the Golgi and increases its expression on the cell surface6.

It has been briefly reported that the F3 3′-UTR contains an inverted Alu element7 (Fig. 1). Alu elements are primate specific, short interspersed mobile DNA sequences8. They first appeared during primate evolution and are now recognised as potential modulators of gene regulation9. They can influence transcriptional and post-transcriptional processes through several mechanisms, including serving as alternative promoters, enhancers, and transcription factor binding sites, or altering DNA methylation patterns9,10,11,12,13. Differential expression of genes with Alu element insertions may play a role in embryo development14. Furthermore, Alu element insertion can lead to protein truncation and thereby alter function15. The functional importance and possible organismal advantage of the Alu element in F3 3′-UTR has not previously been addressed.

This study tested whether the inverted Alu element affects the subcellular localisation of human F3.

Fig. 1
figure 1

UCSC Genome Browser visualisation of the F3 gene highlighting the location of the Alu element. UCSC Genome Browser snapshot of the F3 gene on the human genome assembly (GENCODE v48, hg38). The GENCODE Genes track (top) illustrates the exon–intron structure of the gene and 3′-UTR, and the RepeatMasker track (bottom) identifies Short Interspersed nuclear elements (SINE), including the Alu element (highlighted by the red box).

Full size image

Materials and methods

Cell lines

The MDA-MB-231 (a human breast cancer cell line that expresses F3 constitutively16 cell line were purchased from the ECACC.

Confocal microscopy constructs

Fluorescent F3 constructs (FC1, FC2, FC3, and FC4) were subcloned into the pcDNA™3.1/Zeo(+) vector by GeneArt. GFP-EEA1 wt was a gift from Silvia Corvera (Addgene plasmid # 42307; http://n2t.net/addgene:42307; RRID: Addgene_42307)17; mEmerald-Nucleus-7 was a gift from Michael Davidson (Addgene plasmid # 54206; http://n2t.net/addgene:54206; RRID: Addgene_54206, unpublished); PA-GFP Golgi was a gift from Michael Davidson (Addgene plasmid # 57164; http://n2t.net/addgene:57164; RRID: Addgene_57164, unpublished); GFP- SEC61B was a gift from Christine Mayr (Addgene plasmid # 121159; http://n2t.net/addgene:121159; RRID: Addgene 121159)18.

Transfections

On the day of transfection, MDA-MB-231 cells were washed with PBS. Cells were then incubated in serum-free media (DMEM media with 1% L-glutamine; 500 µl) at 37 °C and 5% CO2 for 30 min. Opti-MEM™ (50 µl) was added to two wells in a 96-well plate per transfection reaction. 2 µl of Lipofectamine™ 2000 was added to one well of Opti-MEM™, and the 1 µg of DNA added to the other. The DNA-Opti-MEM™ was then added to the Lipofectamine™-Opti-MEM™ dropwise and left to incubate at room temperature for 10 min to form the DNA-Lipofectamine™ complex. The complex (100 µl) was then carefully added to the cells, after which the plate gently shaken to cover the cells and incubated at 37 °C and 5% CO2. After 5 h, 500 µl full DMEM media with 1% L-glutamine and 10% FBS was added to the cells and further incubated at 37 °C and 5% CO2, in a humidified tissue culture incubator.

Confocal microscopy

MDA-MB-231 cells were seeded at a concentration of 5 × 104 in 4-well glass bottomed slides, in base media, 24 h prior to transfection. Cells were washed with PBS and incubated at 37 °C and 5% CO2 in serum free media for 30 min. 0.5 µg of DNA was transfected into cells using Opti-MEM™ and Lipofectamine™ 2000. The cells were left at 37 °C and 5% CO2 for 5 h, when full base DMEM media was added.

CellMask™ Orange was used as a cell membrane marker. It was diluted to 2x working solution in pre-warmed base media and added to cells 30 min prior to imaging. Cells were incubated at 37 °C and 5% CO2.

The PAR-2 agonist peptide SLIGKV-NH₂ (R&D Systems) was reconstituted according to the manufacturer’s instructions and used at a final concentration of 100 µM. Cells were treated for the time period specified in the Figure legends. 24 h following transfection, media was changed for fresh base media. 0.5 µl of working solution Hoechst 33342 nuclear stain was added to each well, and the cells covered in foil to protect from the light and left for 10 min. Media was then carefully removed and replaced with FluoroBrite™ DMEM media ready for imaging.

Cells were imaged using a Zeiss LSM-780 inverted confocal microscope. The objective lens used is specified in the Figure legend. For live cell imaging, the chamber was heated to 37 °C with 5% CO2 conditions.

ImageJ software was used to visualise and analyse the images. An EzColocalization ImageJ plugin developed by Stauffer W., et al. was used to conduct all the immunofluorescence colocalization analysis19.

Results

Effects of the Alu insertion on F3 post-translational protein distribution

Fluorescent reporter constructs were designed with or without the Alu element in the human F3 3′-UTR (Fig. 2a) and expressed in cells to track protein localisation. Both constructs colocalise equally intracellularly and extracellularly if they have the same 3′-UTR (Fig. 2b and c), confirming neither fluorophore influences localisation. However, when constructs with different 3′-UTRs are dual-transfected (either mCh-WT/mNG-∆Alu or mNG-WT/mCh-∆Alu), the protein generated via the ∆Alu 3′-UTR showed more trafficking to the cell surface than that generated from the construct with the WT 3′-UTR (Fig. 2d and e). The overall distribution of mCh-WT F3 3′-UTR showed a similar distribution to endogenous F3 protein (Figure S1).

Fig. 2
figure 2

Co-transfection of WT and Δ Alu fluorescent constructs. (a) The top row shows the WT human F3 3′-UTR construct. For each of the fluorescent constructs the ectodomain has been replaced by either mCherry or mNeonGreen. The Alu element is highlighted by a red box in the 3′-UTR, and red dotted line indicates a deletion of the Alu element; (b–e) Fluorescent constructs were co-transfected into MDA-MB-231 cells. Images were acquired as a Z-stack using Plan-Apochromat 63x/1.40 Oil DIC M27 lens on Zeiss LSM-780 inverted confocal microscope. Representative of multiple images, shown as a Z-project. Scale bars, 5 μm. Images analysed on Fiji software (ImageJ). (b) mCh-WT and mNG-WT; (c) mCh-ΔAlu and mNG-ΔAlu; (d) mNG-WT and mCh-ΔAlu; and (e) mCh-WT and mNG-ΔAlu. Fluorophores encoded by the ΔAlu 3′-UTR F3 constructs localise to the cell surface more readily than those encoded by the WT 3’-UTR F3 constructs.

Full size image

We next looked in more detail at differential localisation using a panel of membrane and organelle markers and Pearson correlation analysis. ∆Alu constructs showed significantly greater colocalisation with the plasma membrane (Fig. 3a, Figure S2) and early endosomes (Fig. 3b, Figure S3) compared to WT 3′UTR constructs. Conversely, colocalisation of mCh-∆Alu with a Golgi marker was significantly less than the colocalisation of mCh-WT (Fig. 3c, Figure S4). There was no correlation of either mCh-WT or mCh-∆Alu with markers of the nucleus (Fig. 3d, Figure S5) or endoplasmic reticulum (Fig. 3e, Figure S6). Although to the naked eye the differences between constructs are subtle, significance is shown by the image analysis. Overall, this indicates that fluorescent proteins generated from constructs without the 3’-UTR Alu element (i.e. mCh-∆Alu and mNG-∆Alu) are more readily trafficked to the cell surface than those generated from constructs containing the 3’-UTR Alu element.

Fig. 3
figure 3

The presence of the Alu element in the F3 3′-UTR affects its localisation within MDA-MB-231 cells. MDA-MB-231 cells were co-transfected with markers to identify specific organelles along with a fluorophore under the influence of the F3 3′-UTR with either the Alu element present (WT) or not (∆Alu). (a) cell membranes, (b) early endosomes, (c) Golgi, (d) nucleus, (e) endoplasmic reticulum. (i) Left columns represent organelle localisation; middle column, WT or ΔAlu F3; and the right column, the merged image (colocalisation) of F3, the organelle marker, and Hoechst 33342 to identify the nucleus (blue). Images were acquired as a Z-stack using Plan-Apochromat 63x/1.40 Oil DIC M27 lens on Zeiss LSM-780 inverted confocal microscope. Representative of multiple images. Scale bars, 5 μm. (ii) Pearson’s correlation coefficient of the colocalisation of the organelle marker with the fluorophore protein under the influence of either the WT or ΔAlu 3′-UTR are shown to the right. The number of cells analysed/the number of independent experiments is indicated on the x axis. Data are expressed as mean ± SEM and analysed using one-way ANOVA. *p < 0.05 **p < 0.01.

Full size image

Effects of PAR-2 stimulation

PAR2-AP (SLIGKV-NH2) is known to mobilise F3 from the Golgi to the cell surface in fibroblasts6. Stimulation of MDA-MB-231 cells with PAR2-AP significantly increased the colocalisation of the mNG-WT with the cell surface (Fig. 4a). As the loss of the Alu insertion reduced fluorophore colocalisation with the Golgi, we asked whether the Alu element affected the response to the PAR2-AP in our experimental system. Whilst there was a significant increase of mNG-WT with CellMask™ following PAR2-AP stimulation (p = 0.0381), a corresponding increase was not seen for mNG-∆Alu (Fig. 4b).

Fig. 4
figure 4

PAR2-AP increases F3 protein to the cell membrane but has no effect when there is no Alu element present in the F3 3′-UTR. (a) Fluorescence confocal microscopy of live MDA-MB-231 cells after transfection of mNG-WT (green, left panels), following stimulation with PAR2-AP (0, 1, 2, or 4 h) and stained with CellMask™ Orange visualising plasma membrane (red), and Hoechst 33,342 (blue). (i) The left panel shows plasma membrane, the middle column shows the fluorescent F3 constructs, and the right panel shows a merged image. Colocalisation of F3 with plasma membrane appears yellow. PAR2-AP increases F3 expression in the cell membrane. Images were acquired as a Z-stack using Plan-Apochromat 63x/1.40 Oil DIC M27 lens on Zeiss LSM-780 inverted confocal microscope. Representative of multiple images. Scale bars, 5 μm. (ii) Pearson’s correlation coefficient of the colocalisation of the plasma membrane with mNeonGreen protein under the influence of F3′s WT 3′-UTR are shown. The number of cells analysed/the number of independent experiments is indicated on the x axis. Data are expressed as mean ± SEM and analysed using one-way ANOVA. *p < 0.05; (b) Similar confocal microscopy after transfection of mNG-WT (green) or mNG-ΔAlu (green), following stimulation with PAR2-AP for 4 h and stained with CellMask™ Orange visualising the plasma membrane (red), and DAPI (blue). (i) The left panel shows plasma membrane, the middle column shows the fluorescent F3 constructs, and the right panel shows a merged image. Images were acquired as a Z-stack using Plan-Apochromat 63x/1.40 Oil DIC M27 lens on Zeiss LSM-780 inverted confocal microscope. Representative of multiple images. Scale bars, 5 μm. (ii) Pearson’s correlation coefficient of the colocalisation of the plasma membrane with mNeonGreen protein under the influence of F3′s WT and ΔAlu 3′-UTR are shown. The number of cells analysed/the number of independent experiments is indicated on the x axis. Data are expressed as mean ± SEM and analysed using one-way ANOVA. *p < 0.05.

Full size image

Discussion

We have shown that removal of the inverted Alu element from the F3 3′-UTR affects the subcellular localisation of F3 protein. Using fluorophore reporter constructs, we found that in the absence of the Alu element there was increased plasma membrane and early endosomal localisation of the reporter and reduced presence in the Golgi. Human F3 is known to localise to the Golgi, and to mobilise to the plasma membrane following cell activation by PAR-2 receptor ligation5. Our findings suggest that the Alu element may influence intracellular trafficking, potentially promoting Golgi retention of F3.

PAR-2 is a G-protein-coupled receptor involved in various physiological processes such as coagulation, inflammation, and tissue repair20. It is known to trigger F3 mobilisation from the Golgi to the cell surface upon activation by proteases potentially present during tissue injury5. Consistent with this, we observed increased surface expression of the fluorophore upon PAR-2 activation, but only if the Alu element is present in the F3 3′-UTR. This supports the hypothesis that the Alu insertion enables enhanced Golgi storage of F3, allowing for rapid mobilisation in response to cellular cues. Such rapid mobilisation of an intracellular pool may be advantageous during acute injury or inflammation, enabling swift haemostatic responses.

One possibility is that the Alu may affect mRNA localisation and subsequent translation, a process that mostly takes place in the endoplasmic reticulum, but which can also take place in the Golgi21. Alternatively, there is the possibility that the Alu insertion in 3′-UTR may affect post-translational protein trafficking by affecting the assembly of proteins on F3 during or after its translation. A precedent for such a mechanism comes from the work of Berkovits et al., which showed that alternative poly-adenylation of CD47 mRNA alters the assembly of proteins on CD47 during or after its translation. This affects CD47 protein localisation to the plasma membrane without affecting mRNA abundance, isoform levels, or protein translation efficiency22. A further consideration is that the Alu element in the F3 3′UTR might act as a binding site for a non-coding RNA(s) that affects mRNA and/or protein localisation23.

In conclusion, our findings reveal that the Alu element in the F3 3′-UTR significantly influences the subcellular trafficking of F3 protein, likely enhancing its readiness for mobilisation during haemostatic or inflammatory challenges. This suggests that the primate-specific evolutionary insertion has functional consequences that may provide a selective advantage. The presence of the inverted Alu element in the F3 mRNA 3′-UTR could possibly enhance the dual function of F3 in coagulation and infection control. More generally, altered intracellular trafficking should be added to the possible effects of Alu insertions on gene expression.