Pivotal Role of Translokin/CEP57 in the Unconventional Secretion Versus Nuclear Translocation of FGF2

Intracellular trafficking of fibroblast growth factor 2 (FGF2) exhibits two unusual features: (i) it is secreted despite the lack of signal peptide and (ii) it can translocate to the nucleus after interaction with high- and low-affinity receptors on the cell surface, although it does not possess any classical nuclear localization signal. This nuclear translocation constitutes an important part of the response to the growth factor. Previously, we identified Translokin/CEP57, an FGF2 binding partner, as an intracellular mediator of FGF2 trafficking, which is essential for the nuclear translocation of the growth factor. Here, we report the identification of four Translokin partners: sorting nexin 6, Ran-binding protein M and the kinesins KIF3A and KIF3B. These proteins, through their interaction with Translokin, are involved in two exclusive complexes allowing the bidirectional trafficking of FGF2. Thus, Translokin plays a pivotal role in this original mechanism. In addition, we show that FGF2 secretion is regulated by a negative loop, retro-controlled by FGF receptor and involving FGF2 itself.     

Nucleolar association of fibroblast growth factor 3 via specific sequence motifs has inhibitory effects on cell growth.

The dual subcellular fate of fibroblast growth factor 3 (FGF3) is determined by the competing effects of amino-terminal signals for nuclear localization and secretion (P. Kiefer, P. Acland, D. Pappin, G. Peters, and C. Dickson, EMBO J. 13:4126-4136, 1994). Mutation analysis has implicated additional basic domains in the carboxy-terminal region of the protein as necessary for nuclear uptake and the association of FGF3 with the nucleoli. Immunogold electron microscopy shows that FGF3 is predominantly within the dense fibrillar component of the nucleolus. A form of FGF3 that localizes exclusively in the nucleus and nucleolus was generated by removing signals for secretion, and expression of this nonsecreted FGF3 in a mammary epithelial cell line resulted in slowly growing colonies of enlarged cells. Thus, nuclear import and nucleolar association of FGF3 are determined by the concerted interaction of several distinct motifs, and the exclusive production of the nuclear isoform can inhibit DNA synthesis and cell proliferation.     

Three forms of rat basic fibroblast growth factor are made from a single mRNA and localize to the nucleus

The molecular weight of rat basic fibroblast growth factor is predicted to be 18 kDa when the amino acid sequence is read from the single AUG initiation codon found in the cDNA. DNA sequencing upstream of this AUG codon indicated, however, that there was an extended open reading frame. In vitro translation of the rat cDNA for basic FGF gave three proteins of 18.0, 21.5, and 22.0 kDa in equal abundance. The same proteins were produced in vivo by COS cells transfected with the rat cDNA. Deletion of 81 base pairs from the reading frame upstream of the AUG codon resulted in the expression of only one protein observed at 18.0 kDa. These results indicated that the 22.0 and 21.5 kDa forms of rat basic FGF were formed when translation initiates at the alternative upstream non-AUG codons. Rat cell lines and tissues were found to express all three forms of basic FGF protein. The cDNA was used to analyze the subcellular distribution of the different forms of rat basic FGF. Subcellular fractionation and immunofluorescence of transfected COS cells showed that all three forms of the protein localized preferentially in the nucleus. Expression of a truncated cDNA from which 81 base pairs (27 amino acids) of the upstream reading frame had been deleted, showed localization of the smaller form of bFGF alone in the nucleus. These results demonstrated that although the amino acids that were deleted from the N-terminus of rat basic fibroblast growth factor have a sequence characteristic of nuclear localization motifs, they are not obligatory for the transport of the growth factor into the nucleus. Nuclear extracts taken from transfected cells also contained two smaller proteins of 16 and 12 kDa that were detected by Western blot analysis. It is possible that these are proteolytic products of bFGF.     

Control of the intracellular signaling induced by fibroblast growth factors (FGF) over the proliferation and survival of retinal pigment epithelium cells: example of the signaling regulation of growth factors endogenous to the retina ]

Retinal pigmented epithelium (RPE) cells are of central importance in the maintenance of neural retinal function. RPE cell apoptosis is responsible for the development of a variety of retinal degeneration. The role of FGF2 was investigated on RPE cell proliferation and apoptosis in vitro. In the absence of serum, RPE cells died by apoptosis, while the addition of FGF2 greatly reduces apoptosis over a 7-day culture period. This is due to an autocrine loop involving secretion of endogenous FGF1 in the mechanism that govern FGF2-induced resistance to apoptosis. FGF2 induces long-term activation of FGFR1 and ERK1/2, and production of the anti-apoptotic protein BcL-x. Because FGF1 has no classical signal sequence to direct its secretion, we investigated the effects of FGF1 secretion on RPE proliferation and apoptosis in the absence of exogenous FGF2. Forced secretion of endogenous FGF1 by adding a signal peptide to the FGF1 molecule induces FGF1 secretion and cell proliferation in the presence of serum, while in FGF1 stops to be secreted and cell die in the absence of serum. Conversely, in cells cultured in the presence of serum, FGF1 without signal peptide is not secreted, but is secreted and rescue RPE cell from apoptosis when cells are cultured without serum. Thus, the proliferation and survival activities of endogenous FGF1 depend on the secretion of FGF1 which is determined by the cell environment.     

Retention of fibroblast growth factor 3 in the Golgi complex may regulate its export from cells.

The fibroblast growth factors (FGFs) fall into two distinct groups with respect to their mode of release from cells. Whereas FGF1 and FGF2 lack conventional signal peptides, the remaining members have typical features of secreted proteins. However, the behavior of mouse FGF3 is anomalous, since, despite entering the secretory pathway and undergoing primary glycosylation, its release from transfected COS-1 cells is very inefficient compared with that of FGF4 and FGF5. To investigate the unusual properties of FGF3, we analyzed the processing, secretion, and intracellular localization of a series of site-directed mutants as well as chimeras produced by fusing parts of FGF3, FGF4, and FGF5. Wild-type FGF3 was shown to accumulate in an immature form in the Golgi complex, from where it is slowly released into the extracellular matrix. Removing or relocating the Asn-linked glycosylation site further impaired its release, and exchanging the signal peptide or carboxy terminus had little effect. In contrast, a chimeric protein with an amino terminus from FGF5 was efficiently secreted and biologically active in cell transformation assays. The data suggest that a structural feature of FGF3 involving the amino-terminal region and glycosylation site has a significant bearing on its passage through the Golgi complex and may regulate the secretion of the ligand.     

Intracellular trafficking of LET-756, a fibroblast growth factor of C. elegans, is controlled by a balance of export and nuclear signals

The superfamily of fibroblast growth factors (FGF), which counts 22 members in humans, exerts many functions during animal development and adult life. LET-756 is one of the two FGFs of the nematode C. elegans. Re-introduction of LET-756 in a null mutant strain restores viability, allowing the study of structural requirements for LET-756 trafficking and function. LET-756 protein has several regions and motifs, including a non-classical internal motif required for secretion. We show here that a main difference in the wild-type LET-756 molecule and a truncated molecule that mimics a partial loss-of-function mutant lies on subnuclear expression. Using Cos-1 cells and rescue activity we show that: (i) nuclear localization is due to various redundant NLS, one of them acting as a nucleolar localization signal; (ii) nuclear LET-756 is addressed to the speckles by a stretch of glutamine residues; (iii) nuclear LET-756 is trafficking between speckles and nucleoli; (iv) in the nucleolus, LET-756 is associated with proteins of the rRNA splicing compartment; (v) changing LET-756 secretion signal prevents its nuclear localization. We propose that LET-756 exerts its functions through a balance between secreted and nuclear forms due to two opposite addressing signals, (i) synergy of several NLS and (ii) attenuated secretion signal.     

Nucleotide sequence of the membrane-bound aldehyde dehydrogenase gene from Acetobacter polyoxogenes

The nucleotide sequence of the membrane-bound aldehyde dehydrogenase (ALDH) gene from an industrial vinegar producer, Acetobacter polyoxogenes, was determined. Comparison of the sequence with the NH2-terminal amino acid sequence of the mature ALDH and determination of the actual translational initiation codon by means of in vitro manipulation of the upstream and proximal regions of the cloned gene showed that ALDH was primarily translated as a 773-amino-acid protein and that the 44-amino-acid sequence at the NH2-terminus, which probably serves as a signal peptide, was processed during maturation and localization in the membrane. When ALDH was expressed in a large quantity in Escherichia coli cells after the coding region had been placed downstream of the lac promoter, the ALDH protein, which still contained the signal peptide and had no ALDH activity, was localized in the membrane fraction.     

Fibroblast growth factor 3, a protein with a dual subcellular fate, is interacting with human ribosomal protein S2

The secreted isoform of fibroblast growth factor 3 (FGF3) induces a mitogenic cell response, while the nuclear form inhibits cell proliferation. Recently, we identified a nucleolar FGF3-binding protein which is implicated in processing of pre-rRNA as a possible target of nuclear FGF3 signalling. Here, we report a second candidate protein identified by a yeast two-hybrid screen for nuclear FGF3 action, ribosomal protein S2, rpS2. Recombinant rpS2 binds to in vitro translated FGF3 and to nuclear FGF3 extracted from transfected COS-1 cells. Characterization of the FGF3 binding domain of rpS2 showed that both the Arg-Gly-rich N-terminal region and a short carboxyl-terminal sequence of rpS2 are necessary for FGF3 binding. Mapping the S2 binding domains of FGF3 revealed that these domains are important for both NoBP and rpS2 interaction. Transient co-expression of rpS2 and nuclear FGF3 resulted in a reduced nucleolar localization of the FGF. These findings suggest that the nuclear form of FGF3 inhibits cell proliferation by interfering with ribosomal biogenesis.     

SBP,SECIS binding protein,binds to the RNA fragment upstream of the Sec UGA codon in glutathione peroxidase mRNA

In mammals, most of the selenium contained in the body is present as an unusual amino acid, selenocysteine (Sec), whose codon is UGA. Because the UGA codon is typically recognized as a translation stop signal, it is intriguing how a cell recognizes and distinguishes a UGA Sec codon from a UGA stop codon. For eukaryotic selenoprotein mRNAs, it has been proposed that a conserved stem-loop structure designated the Sec insertion sequence (SECIS) in the 3'-untranslated (3'-UTR) region is required for recognition of UGA as a Sec codon. Some proteins which bind to SECIS (SBP) have been reported. However, it is not clear how the SECIS element in the 3'-UTR can mediate Sec insertion far at the in-frame UGA Sec codons. The idea that there must be a signal near the UGA Sec codon is still considered. Therefore, we searched for a protein which binds to an RNA sequence surrounding the UGA Sec codon on human glutathione peroxidase (GPx) mRNA. We found a protein which strongly bound to the RNA fragment upstream of the UGA Sec codon. However, this protein did not bind to the RNA sequence downstream of the UGA codon. This protein also bound to the SECIS sequence in the 3'-UTR of human GPx, and this binding to SECIS was competed with the RNA fragment upstream of the UGA Sec codon. Comparison of the RNA fragment with the SECIS fragment identified the conserved regions, which appeared in the region upstream of the in-frame UGA Sec codon of Se-protein mRNAs. Thus, this study proposes a novel model to understand the mechanisms of Sec incorporation at the UGA Sec codon, especially the regions upstream of the UGA codon of mRNAs of mammalian selenoproteins. This model explains that the stem-loop structure covering the UGA codon is recognized by SBP and how the UGA Sec codon escapes from attack by eRF of the peptide releasing factor.     

The complete nucleotide sequence of the Pseudomonas gene coding for carboxypeptidase G2

The complete nucleotide sequence of the Pseudomonas chromosomal gene coding for the enzyme carboxypeptidase G2 (CPG2) has been determined. The nucleotide sequence obtained has been confirmed by comparing the predicted amino acid sequence with that of randomly derived peptide fragments and by N-terminal sequencing of the purified protein. The gene has been shown to code for a 22 amino acid signal peptide at its N-terminus which closely resembles the signal peptides of other secreted proteins. An alternative 36 amino acid signal peptide which may function in Pseudomonas has also been identified. The codon utilisation of the gene is influenced by the high G + C (67.2%) content of the DNA and exhibits a 92.8% preference for codons ending in G or C. This unusual codon preference may contribute to the generally observed weak expression of Pseudomonas genes in Escherichia coli. A region of DNA upstream of the structural gene has also been sequenced and a ribosome binding site and two putative promoter sequences identified.     

Effect of the thermostable protein kinase inhibitor on intracellular localization of the catalytic subunit of cAMP-dependent protein kinase.

cAMP-dependent protein kinase mediates a variety of cellular responses in most eukaryotic cells. Many of these responses are cytoplasmic, whereas others appear to require nuclear localization of the catalytic subunit. In order to understand further the molecular basis for subcellular localization of the catalytic subunit, the effect of the heat stable protein kinase inhibitor (PKI) was investigated. The subcellular localization of the catalytic (C) subunit was determined both in the presence and absence of PKI, by microinjecting fluorescently labeled C subunit into single living cells. When injected alone, a significant fraction of the dissociated C subunit localized to the nucleus. When coin-injected with an excess of PKI, little of the C subunit localized to the nucleus, suggesting that accumulation of catalytic subunit in the nucleus requires either enzymatic activity or a nuclear localization signal. Inactivation of the catalytic subunit in vitro by treatment with N-ethylmaleimide did not prevent localization in the nucleus, indicating that enzymatic activity was not a prerequisite for nuclear localization. In an effort to search for a specific signal that might mediate nuclear localization, a complex of the catalytic subunit with a 20-residue inhibitory peptide derived from PKI (PKI(5-24)) was microinjected. In contrast to intact PKI, the peptide was not sufficient to block nuclear accumulation. In the presence of PKI(5-24), the C subunit localized to the nucleus in a fashion analogous to that of dissociated, active C subunit despite evidence of no catalytic activity in situ. Thus, nuclear localization of the C subunit appears to be independent of enzymatic activity but most likely dependent upon a signal. The signal is apparently masked by both the regulatory subunit and PKI but not by the inhibitory peptide.     

Effect of codon-optimized E. coli signal peptides on recombinant Bacillus stearothermophilus maltogenic amylase periplasmic localization,yield and activity

Recombinant proteins can be targeted to the Escherichia coli periplasm by fusing them to signal peptides. The popular pET vectors facilitate fusion of target proteins to the PelB signal. A systematic comparison of the PelB signal with native E. coli signal peptides for recombinant protein expression and periplasmic localization is not reported. We chose the Bacillus stearothermophilus maltogenic amylase (MA), an industrial enzyme widely used in the baking and brewing industry, as a model protein and analyzed the competence of seven, codon-optimized, E. coli signal sequences to translocate MA to the E. coli periplasm compared to PelB. MA fusions to three of the signals facilitated enhanced periplasmic localization of MA compared to the PelB fusion. Interestingly, these three fusions showed greatly improved MA yields and between 18- and 50-fold improved amylase activities compared to the PelB fusion. Previously, non-optimal codon usage in native E. coli signal peptide sequences has been reported to be important for protein stability and activity. Our results suggest that E. coli signal peptides with optimal codon usage could also be beneficial for heterologous protein secretion to the periplasm. Moreover, such fusions could even enhance activity rather than diminish it. This effect, to our knowledge has not been previously documented. In addition, the seven vector platform reported here could also be used as a screen to identify the best signal peptide partner for other recombinant targets of interest.     

Formation of Disulfide Bridges Drives Oligomerization,Membrane Pore Formation,and Translocation of Fibroblast Growth Factor 2 to Cell Surfaces

Fibroblast growth factor 2 (FGF2) is a key signaling molecule in tumor-induced angiogenesis. FGF2 is secreted by an unconventional secretory mechanism that involves phosphatidylinositol 4,5-bisphosphate-dependent insertion of FGF2 oligomers into the plasma membrane. This process is regulated by Tec kinase-mediated tyrosine phosphorylation of FGF2. Molecular interactions driving FGF2 monomers into membrane-inserted FGF2 oligomers are unknown. Here we identify two surface cysteines that are critical for efficient unconventional secretion of FGF2. They represent unique features of FGF2 as they are absent from all signal-peptide-containing members of the FGF protein family. We show that phosphatidylinositol 4,5-bisphosphate-dependent FGF2 oligomerization concomitant with the generation of membrane pores depends on FGF2 surface cysteines as either chemical alkylation or substitution with alanines impairs these processes. We further demonstrate that the FGF2 variant forms lacking the two surface cysteines are not secreted from cells. These findings were corroborated by experiments redirecting a signal-peptide-containing FGF family member from the endoplasmic reticulum/Golgi-dependent secretory pathway into the unconventional secretory pathway of FGF2. Cis elements known to be required for unconventional secretion of FGF2, including the two surface cysteines, were transplanted into a variant form of FGF4 without signal peptide. The resulting FGF4/2 hybrid protein was secreted by unconventional means. We propose that the formation of disulfide bridges drives membrane insertion of FGF2 oligomers as intermediates in unconventional secretion of FGF2.     

Chromatin compaction and cell death by high molecular weight FGF-2 depend on its nuclear localization, intracrine ERK activation, and engagement of mitochondria

Fibroblast growth factor 2 (FGF-2) is produced as CUG-initiated, 22-34 kDa or AUG-initiated 18 kDa isoforms (hi- and lo-FGF-2, respectively), with potentially distinct functions. We report that expression of hi-FGF-2 in HEK293 cells elicited chromatin compaction preceding cell death with apoptotic features. Nuclear localization of the intact protein was required as expression of a non-nuclear hi-FGF-2 mutant failed to elicit chromatin compaction. Equally ineffective, despite nuclear localization, was the over-expression of the 18 kDa core sequence (lo-FGF-2). Chromatin compaction by hi-FGF-2 was accompanied by increased cytosolic cytochrome C, and was attenuated either by over-expression of Bcl-2 or by a peptide inhibitor of the pro-apoptotic protein Bax. In addition hi-FGF-2 elicited sustained activation of total and nuclear extracellular signal regulated kinase (ERK1/2) by an intracrine route, as it was not prevented by neutralizing anti-FGF-2 antibodies. Inhibition of the ERK1/2 activating pathway by dominant negative upstream activating kinase, or by PD 98059, prevented chromatin compaction by hi-FGF-2. ERK1/2 activation was not affected by the Bax-inhibiting peptide suggesting that it occurred upstream of mitochondrial involvement. We conclude that the hi-FGF-2-induced chromatin compaction and cell death requires its nuclear localization, intracrine ERK1/2 activation and mitochondrial engagement.     

Nonsecreted insulin-like growth factor binding protein-3 (IGFBP-3) can induce apoptosis in human prostate cancer cells by IGF-independent mechanisms without being concentrated in the nucleus

Insulin-like growth factor binding protein-3 (IGFBP-3), a secreted protein, has the intrinsic ability to induce apoptosis directly without binding insulin-like growth factors. Previous studies suggested that IGFBP-3 must be secreted to exert its biological functions. IGFBP-3 contains a nuclear localization signal (NLS), and exogenous IGFBP-3 is translocated into the nucleus, suggesting that both secretion and nuclear localization may play important roles in IGFBP-3 action. To address these questions, we fused yellow fluorescent protein (YFP) to mature IGFBP-3 lacking its signal peptide so that it would remain intracellular and mutated the C-terminal NLS of IGFBP-3, (228)KGRKR(232), to MDGEA. Following transfection of PC-3 human prostate cancer cells with these constructs, Western blots indicated that YFP-IGFBP-3 lacking a signal peptide was cell-associated and not present in the extracellular media. Moreover, the fusion protein was not N-glycosylated, indicating that it had not entered the secretory pathway. Confocal imaging showed that intracellular YFP-MDGEA-IGFBP-3 was predominantly cytoplasmic. Transient transfection of nonsecreted YFP-wild-type IGFBP-3 decreased cell viability, as assessed by staining with annexin V followed by flow cytometry. Induction of cell death was caspase-dependent, indicative of apoptosis. Apoptosis also was induced by the nonsecreted NLS mutant (YFP-MDGEA-IGFBP-3) alone and when the IGF-binding site also had been mutated. These results indicate that IGFBP-3 can induce apoptosis in an IGF-independent manner without being secreted or concentrated in the nucleus.     

Functional analysis of a mutation occurring between the two in-frame AUG codons of human angiotensinogen

Angiotensinogen (ANG) is the specific substrate of the renin-angiotensin system, a major participant in blood pressure control. We have identified a natural mutation at the -30 amino acid position of the angiotensinogen signal peptide, in which an arginine is replaced by a proline (R-30P). Heterozygous individuals with R-30P showed a tendency to lowered plasma angiotensinogen level (1563 ng of ANG I/ml (range 1129-1941)) compared with normal individuals in the family (1892 ng of ANG I/ml (range 1603-2072)). Human angiotensinogen mRNA has two in-phase translation initiation codons (AUG) starting upstream 39 and 66 nucleotides from the cap site. R-30P occurs in a cluster of basic residues adjacent to the first AUG codon that may affect intracellular sorting of the nascent protein. Pulse-chase experiments in transiently transfected cultured cells revealed that the R-30P mutation was associated with reduced amounts of both intra- and extracellular protein. In a cell-free system, we found that two forms of native angiotensinogen were generated by alternative initiation of translation at either AUG codon. Alteration of either the first or second AUG codons abolished the synthesis of the longer and the shorter form of native angiotensinogen, respectively. Furthermore, the rate of secretion of the shorter form was lower than that of the longer form. By transplanting angiotensinogen signal peptide onto green fluorescence protein, however, we found that both forms of the signal peptide could target green fluorescence protein, normally localized in the cytoplasm, to the secretory pathway. Although the R-30P mutation may not affect intracellular sorting of angiotensinogen in a qualitative manner, it leads to a quantitative reduction in the net secretion of mature angiotensinogen through decreased translocation or increased residence time in the endoplasmic reticulum.     

The extended signal peptide of the trimeric autotransporter EmaA of Aggregatibacter actinomycetemcomitans modulates secretion

The extracellular matrix protein adhesin A (EmaA) of the Gram-negative bacterium Aggregatibacter actinomycetemcomitans is a fibrillar collagen adhesin belonging to the family of trimeric autotransporters. The protein forms antenna-like structures on the bacterial surface required for collagen adhesion. The 202-kDa protein monomers are proposed to be targeted and translocated across the inner membrane by a long signal peptide composed of 56 amino acids. The predicted signal peptide was functionally active in Escherichia coli and A. actinomycetemcomitans using truncated PhoA and Aae chimeric proteins, respectively. Mutations in the signal peptide were generated and characterized for PhoA activity in E. coli. A. actinomycetemcomitans strains expressing EmaA with the identical mutant signal peptides were assessed for cellular localization, surface expression, and collagen binding activity. All of the mutants impaired some aspect of EmaA structure or function. A signal peptide mutant that promoted alkaline phosphatase secretion did not allow any cell surface presentation of EmaA. A second mutant allowed for cell surface exposure but abolished protein function. A third mutant allowed for the normal localization and function of EmaA at 37°C but impaired localization at elevated temperatures. Likewise, replacement of the long EmaA signal peptide with a typical signal peptide also impaired localization above 37°C. The data suggest that the residues of the EmaA signal peptide are required for protein folding or assembly of this collagen adhesin.     

八肋游仆虫第二类肽链释放因子核定位信号分析

蛋白质合成终止过程中肽链释放因子负责终止密码子的识别.真核生物第二类肽链释放因子（eRF3）是一类GTP酶,协助第一类肽链释放因子（eRF1）识别终止密码子和水解肽酰 tRNA酯键.之前的研究表明,两类肽链释放因子在细胞核中发挥功能,参与蛋白质合成和纺锤体的组装.本研究根据软件预测结果,构建了一系列八肋游仆虫eRF3的截短型突变体,分析在其N端是否存在引导eRF3的核定位信号.结果表明,在eRF3的N端有两个区域（NLS1:23-36 aa 和 NLS2: 236-272 aa）可以引导eRF3进入细胞核中,而且这两个区域具有典型的核定位信号的氨基酸序列特征. eRF3的核定位与其作为一种穿梭蛋白的功能相一致,即参与细胞有丝分裂纺锤体的形成和无义介导的mRNA降解途径.