Expression of transfected vimentin genes in differentiating murine erythroleukemia cells reveals divergent cis-acting regulation of avian and mammalian vimentin sequences

Vimentin expression is retained in erythroid cells differentiated from human iPSC and ESC and indicates dysregulation in these cells early in differentiation

Background

Pluripotent stem cells are attractive progenitor cells for the generation of erythroid cells in vitro as have expansive proliferative potential. However, although embryonic (ESC) and induced pluripotent (iPSC) stem cells can be induced to undergo erythroid differentiation, the majority of cells fail to enucleate and the molecular basis of this defect is unknown. One protein that has been associated with the initial phase of erythroid cell enucleation is the intermediate filament vimentin, with loss of vimentin potentially required for the process to proceed.

Methods

In this study, we used our established erythroid culture system along with western blot, PCR and interegation of comparative proteomic data sets to analyse the temporal expression profile of vimentin in erythroid cells differentiated from adult peripheral blood stem cells, iPSC and ESC throughout erythropoiesis. Confocal microscopy was also used to examine the intracellular localisation of vimentin.

Results

We show that expression of vimentin is turned off early during normal adult erythroid cell differentiation, with vimentin protein lost by the polychromatic erythroblast stage, just prior to enucleation. In contrast, in erythroid cells differentiated from iPSC and ESC, expression of vimentin persists, with high levels of both mRNA and protein even in orthochromatic erythroblasts. In the vimentin-positive iPSC orthochromatic erythroblasts, F-actin was localized around the cell periphery; however, in those rare cells captured undergoing enucleation, vimentin was absent and F-actin was re-localized to the enucleosome as found in normal adult orthrochromatic erythroblasts.

Conclusion

As both embryonic and adult erythroid cells loose vimentin and enucleate, retention of vimentin by iPSC and ESC erythroid cells indicates an intrinsic defect. By analogy with avian erythrocytes which naturally retain vimentin and remain nucleated, retention in iPSC- and ESC-derived erythroid cells may impede enucleation. Our data also provide the first evidence that dysregulation of processes in these cells occurs from the early stages of differentiation, facilitating targeting of future studies.

Electronic supplementary material

The online version of this article (10.1186/s13287-019-1231-z) contains supplementary material, which is available to authorized users.

Adult chicken alpha-globin gene expression in transfected QT6 quail cells: evidence for a negative regulatory element in the alpha D gene region

The chicken adult alpha-globin genes, alpha A and alpha D, are closely linked in chromosomal DNA and are coordinately expressed in vivo in an approximate 3:1 ratio, respectively. When subcloned DNAs containing one or the other gene are stably transfected into QT6 quail fibroblasts, the alpha A-globin gene is expressed at measurable RNA levels, but the alpha D gene is not. The alpha A gene expression can be considerably increased by the presence of a linked Rous sarcoma virus long terminal repeat enhancer, but that of the alpha D gene remains undetectable. Transfection with subclones containing both genes, either in cis or in trans, leads to considerably reduced alpha A RNA levels and still no observable alpha D gene expression. Transfection with deleted subclones suggests that maximal expression levels in this system require the alpha A-globin gene promoter, as opposed to that of the alpha D gene, but that such expression is greatly reduced by one or more DNA sequences which lie approximately 2,000 base pairs upstream of the alpha A gene, within the body of the alpha D gene.

Desmin/vimentin intermediate filaments are dispensable for many aspects of myogenesis

An expression vector was prepared containing a cDNA coding for a truncated version of the intermediate filament (IF) protein desmin. The encoded truncated desmin protein lacks a portion of the highly conserved alpha-helical rod region as well as the entire nonhelical carboxy-terminal domain. When transiently expressed in primary fibroblasts, or in differentiating postmitotic myoblasts and multinucleated myotubes, the truncated protein induces the complete dismantling of the preexisting vimentin or desmin/vimentin IF networks, respectively. Instead, in both cell types vimentin and desmin are packaged into hybrid spheroid bodies scattered throughout the cytoplasm. Despite the complete lack of intact IFs, myoblasts and myotubes expressing truncated desmin assemble and laterally align normal striated myofibrils and contract spontaneously in a manner indistinguishable from that of control myogenic cells. In older cultures the spheroid bodies shift from a longitudinal to a predominantly transverse orientation and loosely align along the I-Z-I-regions of striated myofibrils (Bennett, G.S., S. Fellini, Y. Toyama, and H. Holtzer. 1979. J. Cell Biol. 82:577-584), analogous to the translocation of intact desmin/vimentin IFs in control muscle. These results suggest the need for a critical reexamination of currently held concepts regarding the functions of desmin IFs during myogenesis.

Regulation of the expression of vimentin gene during the differentiation of mouse myeloid leukemia cells

We have examined the expression of vimentin during the differentiation of mouse myeloid leukemia cells (M1), which were induced to differentiate into macrophages by exposure to conditioned medium (CM) obtained from rat embryo fibroblasts. The synthesis of vimentin, which was examined by two-dimensional gel electrophoresis, increased after 12-24 h of incubation of M1 cells in CM and the elevated level of synthesis continued up to 96 h. A macrophage cell line (Mm1) that was derived from spontaneously differentiated M1 cells constantly synthesized much higher levels of vimentin. The amount of vimentin, which was revealed by immunoblot analysis using an mAb against human vimentin, also increased after differentiation by a factor of 7 when compared on the basis of constant protein and by a factor of 17 on the basis of constant cell numbers. Mm1 cells contained greater than 12- and 45-fold more vimentin compared with undifferentiated M1 cells on the bases of constant protein and constant cell numbers, respectively. Northern blot analysis using vimentin cDNA as a probe revealed increases in vimentin mRNA in the differentiated M1 cells and Mm1 cells. Nuclear run-on assay showed that the expression of vimentin gene during the differentiation of M1 cells was transcriptionally regulated. Observations in indirect immunofluorescence microscopy and EM clearly showed that vimentin bundles were rarely observed in undifferentiated M1 cells, and increased amounts of and large-size vimentin bundles were easily observed in differentiated M1 and Mm1 cells. These results suggest the participation of increased amounts of vimentin filaments in the maldistribution of nuclei in M1 cells during differentiation.

Expression of the chicken vimentin gene in transgenic mice: efficient assembly of the avian protein into the cytoskeleton

To study expression and function of the vimentin gene, transgenic mice were generated by microinjecting the entire chicken gene plus 2.4 kilobases of 5' and 2.6 kilobases of 3' flanking sequences. All the transgenic mice obtained had incorporated multiple copies of the gene. RNA analyses demonstrated that the chicken vimentin gene was efficiently expressed in an appropriate tissue-specific pattern and that the transcripts were properly processed, as in chicken, giving rise to two RNAs. The vimentin transgene was predominantly expressed in lens at levels of up to 10-fold the endogenous level in every transgenic line studied. The chicken vimentin transcripts were efficiently translated into polypeptides that were modified posttranslationally and could assemble into the mouse cytoskeleton. Overexpression of the chicken vimentin gene did not obviously affect the expression of the endogenous gene at the RNA or the protein level. Immunofluorescence microscopy further demonstrated that the chicken protein was properly expressed spatially in lens. However, the levels were much higher in the transgenic animals.

Transgenic expression of the muscle-specific intermediate filament protein desmin in nonmuscle cells

The coding region of the hamster desmin gene was fused to the 5' flanking sequences of the hamster vimentin gene and introduced into the germ line of mice. The expression of this intermediate filament gene construct (pVDes) was analyzed at the RNA and protein level in transgenic mice as well as in fibroblast cell lines and primary hepatocyte cultures derived from these mice. In all transgenic mice, the pVDes-encoded protein was coexpressed with mouse vimentin in a tissue-specific fashion and was indistinguishable from normal hamster desmin. Culturing of transgenic hepatocytes induced desmin expression indicating that 3.2 kbp of the vimentin gene 5' region regulates both tissue-specific and tissue culture-induced intermediate filament protein expression. Immunohistochemical staining and double-label immunoelectron microscopy of cultured transgenic fibroblasts showed that the pVDes protein assembled into intermediate filaments which colocalized with the mouse vimentin filaments. Endogenous vimentin RNA levels were not influenced by high-level pVDes expression. The coexpression of desmin and vimentin in nonmuscle cells did not result in detectable developmental, morphological, or physiological abnormalities.

Concentration dependence of transcriptional transactivation in inducible E1A-containing human cells

The adenovirus E1A proteins are essential for the normal temporal activation of transcription from every other adenoviral early promoter. High-level E1A expression in the absence of viral infection would facilitate biochemical studies of E1A-mediated transactivation. Toward this end, we introduced the adenovirus type 2 E1A gene under the control of the murine mammary tumor virus promoter into HeLa cells. Uninduced cells expressed little or no detectable E1A mRNA. Upon induction, mRNA levels accumulated to about 50% of the level observed in 293 cells. The level of E1A expression in these cells could be controlled by varying the concentration of the inducing glucocorticoid. Under these conditions of varying E1A concentrations, it was observed that activation of the E2, E3, and E4 promoters of H5dl312 initiated at the same E1A concentration and that transcription from each promoter increased as the E1A concentration increased. These results indicate that E1A-mediated transactivation is proportional to the concentration of E1A protein. E1A-dependent transcriptional stimulation of the E4 promoter was reproduced in an in vitro transcription system, demonstrating that expression of only the E1A proteins was sufficient to increase the transcriptional activity of nuclear extracts.

Concentration dependence of transcriptional transactivation in inducible E1A-containing human cells

The adenovirus E1A proteins are essential for the normal temporal activation of transcription from every other adenoviral early promoter. High-level E1A expression in the absence of viral infection would facilitate biochemical studies of E1A-mediated transactivation. Toward this end, we introduced the adenovirus type 2 E1A gene under the control of the murine mammary tumor virus promoter into HeLa cells. Uninduced cells expressed little or no detectable E1A mRNA. Upon induction, mRNA levels accumulated to about 50% of the level observed in 293 cells. The level of E1A expression in these cells could be controlled by varying the concentration of the inducing glucocorticoid. Under these conditions of varying E1A concentrations, it was observed that activation of the E2, E3, and E4 promoters of H5dl312 initiated at the same E1A concentration and that transcription from each promoter increased as the E1A concentration increased. These results indicate that E1A-mediated transactivation is proportional to the concentration of E1A protein. E1A-dependent transcriptional stimulation of the E4 promoter was reproduced in an in vitro transcription system, demonstrating that expression of only the E1A proteins was sufficient to increase the transcriptional activity of nuclear extracts.