Translation of the 60-kDa spastin isoform from the AUG in position 259–261 may depend on the migration of the translational machinery until it meets this AUG codon that lies in a better Kozak's sequence context than the first AUG (Figure 1). However, the program UTRScan predicts a secondary RNA structure compatible with the presence of an IRES, immediately upstream of the second AUG, suggesting that the short spastin isoform could be synthesized through a cap-independent mechanism. To test this possibility, we cloned the SPG4 cDNA sequence between the first and second ATG into a widely used dicistronic vector, pRF (construct pRF +4/+258). This vector contains the SV40 promoter directing the expression of a dicistronic RNA encoding the Renilla luciferase as the first cistron and the firefly luciferase as the second. This plasmid was transfected in HeLa and SH-SY-5Y neuroblastoma cell lines and the Renilla and firefly luciferase activities were measured. The construct pRF +4/+258 displayed a high firefly activity in HeLa cells compared with the control empty vector (Figure 2a), consistent with the possibility that the region between the two AUGs in the first exon of SPG4 might contain a functional IRES. This construct was less active in SH-SY-5Y cells.

Although the dicistronic test has been considered the gold standard for testing the existence of functional IRES elements, a major drawback of this approach is that it cannot distinguish between IRES activity and the presence of a cryptic promoter 23. To exclude this possibility, we cloned the same SPG4 sequence into a promoter-less pRF vector (pRFΔP), in which the SV40 promoter has been removed (Figure 2b). Both Renilla and firefly luciferase activities were almost undetectable when the empty pRFΔP vector was transfected, whereas a dramatic increase of firefly activity was observed for the pRFΔP +4/+258 construct in HeLa cells, strongly suggesting the presence of a promoter activity in the first exon of SPG4 (Figure 2b). Again, the fold of activation was lower in SH-SY-5Y cells (Figure 2b).

The presence of a strong promoter in the region under analysis could mask the presence of the IRES, hampering the detection of its functionality. To circumvent this problem, an effective method is direct transfection of the dicistronic RNAs 23. To this end, in vitro-transcribed capped dicistronic mRNAs were transfected into HeLa cells and the activities of both Renilla and firefly luciferases were measured. The firefly activities of both the empty vector and the pRF +4/+258 were barely detectable, while the Renilla luciferase activities were comparable, indicating that the first exon of SPG4 does not contain an IRES element (Figure 2c).

The finding of promoter activity in the region between the two ATGs prompted us to study the regulatory sequences of the SPG4 gene. Bioinformatic analysis of the genomic region upstream of the transcriptional start site (defined as in the reference sequence AB029006) does not identify any TATA box, but detects several CG boxes and a CAAT box in position -597. Furthermore, sequence comparison between the human and mouse SPG4 genomic sequence shows a high degree of sequence conservation in the 5'-UTR of SPG4 and in a region of 400 base pairs (bp) upstream of the putative initiation of transcription, suggesting that this region may contain important regulatory elements (Figure 3a).

To define the minimal genomic region that confers basic expression of the SPG4 gene, we tested the ability of different fragments of the genomic region upstream of the first ATG of the SPG4 gene to drive the expression of the luciferase gene in transiently transfected HeLa, HEK293 and SH-SY-5Y cells (Figure 3b). The activities of these promoters were measured by a luciferase assay and considered as fold of induction in respect to the activity of the empty vector. We did not find any cell-specific difference in the activities of the different fragments in the three cell lines (Figure 3c). The construct that showed higher promoter activity was S -621/-1, which contained the highly conserved, 400 bp-genomic region and the 5'-UTR. Inclusion of an additional 669 bp upstream of this region led to a certain decrease of promoter activity, while removal of a sequence of approximately 220 bp containing the CAAT box (S -400/+3) did not reduce significantly the promoter activity. Deletion of the 5'-UTR and approximately 200 bp upstream of the transcription initiation site (construct S -1290/-424) completely abolished promoter activity, while the removal of only the 5'-UTR (S -400/-206) reduced the basal transcriptional activity (Figure 3c). These experiments identified a region of 400 bp upstream of the first ATG as a minimal promoter region active in all cell lines tested and point to a role of the 5'-UTR to sustain basic ubiquitous SPG4 expression.

The promoter activity observed in the region within the two ATGs in the experiments with the promoter-less pRFΔP vector, as well as the role of the 5'-UTR for basal expression, induced us to examine in detail the potential presence of regulatory sequences in exon 1. We cloned different regions of the first exon of SPG4 upstream of the firefly luciferase gene and tested their promoter activity in all cell lines (Figure 4a). We found a strong promoter activity in both HeLa and HEK293 cells in the region that starts immediately downstream of the putative transcriptional start site (TSS) and include both ATGs (S -207/+259) (Figure 4b). This activity, although decreased, is still present in constructs that contain only the coding region (S +4/+259) or the 5'-UTR (S -207/-1). We define the whole region between the canonical TSS, as defined in public databases, and the AUG in position 259–261, as a cryptic promoter. Notably, the activity of this cryptic promoter appears to display some degree of cell-line specificity, being highly functional in HeLa cells and HEK293 and significantly less in SH-SY-5Y cells (Figure 4b), thus confirming our previous observations with the promoter-less dicistronic vector.

We used the TRANSFAC program to identify binding sites for known transcription factors in the cryptic promoter region. This allowed us to identify two putative Sp1 binding sites that were conserved in the human and mouse genomes (Figure 4a). Site-directed mutagenesis was employed to insert mutations into the upstream, the downstream, or both Sp1 sites in the construct S -207/+259. Transfection of the mutated constructs showed a significant reduction of the cryptic promoter activity in HeLa cells only when both Sp1 sites are mutagenized (Figure 4c).

We previously showed that when spastin is expressed in mammalian cells, two isoforms are produced, starting from the first and second methionine 11. The amount of the shorter isoform increases when the 5'-UTR is included in the construct 11. In vitro transcription-translation experiments suggested that this is largely due to alternative initiation of translation 11. However, our novel findings suggest that transcriptional regulation could contribute to the production of the shorter isoform through the use of the cryptic promoter. To test this possibility, we removed the CMV promoter from a CMV-spastin-GFP construct and analyzed the ability of the cryptic promoter (in this construct represented only by the region +4/+259) to drive the expression of the short spastin isoform after transfection in different cell lines. We found that a short spastin-GFP isoform, with a size consistent with initiation of translation at the second AUG, is produced in this condition in Hela, HEK293 and murine spinal motoneuronal NSC34 cells (Figure 5a and 5b and data not shown). To confirm this data, we generated a construct containing the GFP reporter under the control of the CMV promoter followed by a stop codon and by the coding region of spastin (CMV-EGFP-STOP-Spastin). Such a construct could express spastin only if the region between the first two ATGs functions as a promoter. Consistently, transfected cells with high levels of GFP expression showed low levels of spastin expression, detected with a specific antibody (Figure 5c).

In conclusion, both reporter and expression studies with promoter-less constructs strongly indicate that the first SPG4 exon contains a cryptic promoter that may contribute to produce the 60-kDa isoform in several cell types in vivo.

The S44L and P45Q (c.131C>T and c.134C>A) polymorphisms in the SPG4 gene act as phenotype-modifiers. Patients that bear one of these polymorphisms and a canonical SPG4 mutation on the other allele show an early age onset of HSP and rapid progression of symptoms 1516. Since these nucleotide changes fall into the newly identified cryptic promoter, we tested their capability to affect the promoter activity. The polymorphisms were inserted by mutagenesis in the constructs S -207/+259 and S +4/+259 (Figure 6a). The activity of these mutagenized promoters was tested in HeLa cells and compared with wild-type constructs. The presence of the c.131C>T substitution significantly diminished the activity of the promoter by about a half, while no effect was detected with the c.134C>A substitution (Figure 6b). When the substitutions were inserted in the context of a larger promoter, also containing part of the ubiquitous minimal promoter (S -400/+259), no change in activity was observed for either (Figure 6b).

The previous experiments strongly suggest the existence of a cryptic promoter in the first exon of the SPG4 gene. Differentially regulated, alternative TSSs are a common feature in protein-coding genes and commonly generate alternative N-termini 24. Genome-wide analyses, using short tags derived from the 5'-ends of capped RNAs (CAGE), oligocapping methods and full-length cDNA collections, can be publicly accessed in the CAGE analysis website and in the database of transcriptional start sites (DBTSS). We searched these databases for TSSs within the human SPG4 gene. Remarkably, we found that both databases identify an alternative promoter located within exon 1, downstream of the first ATG, defined by a clustering of TSSs separated by fewer than 500 bp. The tags derive from HEK293 cells, as well as from different tissues, including brain tissue. A summary of these data is represented in Additional file 1 and additional file 2.

To gain further experimental proof that the cryptic promoter is responsible for the synthesis of a short SPG4 mRNA, we performed 5'-end RACE experiments in both HeLa and SH-SY-5Y cells. Total RNA was isolated from the cells. Truncated or uncapped RNA molecules were removed by a phosphatase treatment. Subsequently, the caps were eliminated by treatment with tobacco acid pyrophosphatase (TAP), and an adapter oligonucleotide was ligated to the 5'-ends. After nested amplification with gene-specific primers located downstream of the second AUG, we could amplify in HeLa cells a specific product of about 250 bp, which was absent from the minus TAP control reaction (Figure 7a). This product was cloned and sequenced and found to initiate from nucleotide +117. This transcript therefore, contains an ORF that starts with AUG 259–261 and encodes the short 60-kDa spastin isoform. Both Sp1 sites are located upstream of the beginning of the novel transcript, while the polymorphisms c.131C>T and c.134C>A appear to be positioned a few bases downstream (Figure 7b). Notably, this TSS corresponds to two tags identified in HEK293 in the DBTSS database. We could not obtain a similar 5'-end RACE product in SH-SY-5Y cells, consistent with the lower activity of the cryptic promoter in this cell line (not shown).

In all numeric references in this study, nucleotide +1 corresponds to the A of the first ATG codon according to den Dunnen and Antonarakis 32.

Site-directed mutagenesis was performed by PCR reactions using Pfu Ultra (Stratagene). After amplification, 10 U of DpnI were added to the PCR product and incubated for 1 hour at 37°C. The mutagenized DNA was transformed into E. coli XL1Blue super-competent cells. The c.131C>T polymorphism was introduced into S -400/+259, S -207/+259 and S +4/+259 vectors using the following set of oligos: Fw: 5'-GCCCCTCCGCCCGAGTTGCCGCATAAGCGGAAC-3', Rev: 5'-GTTCCGCTTATGCGGCAACTCGGGCGGAGGGGC-3' (mismatches reported in bold). The c.134C>A polymorphism was introduced into S -400/+259, S -207/+259 and S +4/+259 vectors using the following set of oligos: Fw: 5'-CCTCCGCCCGAGTCGCAGCATAAGCGGAACCTG-3', Rev: 5'-CAGGTTCCGCTTATGCTGCGACTCGGGCGGAGG-3' (mismatches reported in bold). Mutagenesis of Sp1 sites was introduced in the S-207/+259 construct. The upstream Sp1 site was mutated using this set of oligos: Fw: 5'-AGGAAGGAGAAAGGGGAAGGGCAAGCGGGCAGCGTGCGG-3', Rev: 5'-CCGCACGCTGCCCGCTTGCCCTTCCCCTTTCTCCTTCCT-3' (mismatches reported in bold). The downstream Sp1 site was mutated using the following set of oligos: Fw: 5'-CCCTTGCCTGGCCCCAACCCCAACCGCCGCCGGGCCGGC-3', Rev: 5'-GCCGGCCCGGCGGCGGTTGGGGTTGGGGCCAGGCAAGGG-3' (mismatches reported in bold). All mutagenized vectors were controlled by DNA sequencing.

DNA sequencing was performed by using a 3100 Genetic Analyzer (Applied Biosystems) and BigDye Terminator v1.1 Cycle Sequencing kit (Applied Biosystems) according to the manufacturer's specifications.

Renilla (RL) and firefly luciferase (FL) activities were measured using the Dual-Luciferase Reporter System (Promega) and a Victor² 1420 Multilabel Counter (Perkin Elmer). At 24 hours post-transfection, 20 μl of cell lysate was combined sequentially with FL- and RL-specific substrates according to the protocol supplied by the manufacturer. Light emission was measured 2 seconds after addition of each of the substrates and integrated over a 10-second interval. All experiments were performed in duplicates and were repeated at least three times using different DNA preparations.

pRF and pBS-pRF +4/+258 plasmids were linearized prior to transcription by BamHI restriction, purified by incubation at 50°C for 30 minutes with 10 μg proteinase K and 0.5% sodium dodecyl sulfate (SDS), and precipitated with 25 mM ethylene diamine tetraacetic acid (EDTA) and 300 mM sodium acetate pH 5.2. Capped RNA transcripts were synthesized by using MAXIscript in vitro transcription kit (Ambion) according to the manufacturer's specifications. Briefly, recombinant T7 or T3 polymerases were used to synthesize mRNA from 2.5 μg linearized DNA and 0.5 mM Ribo m⁷ G Cap Analog (Promega) was added to the reaction mix. In vitro transcription was performed by incubation at 37°C for 1 hour in 40 U of RNAsin RNAse inhibitor (Promega). Following transcription, reactions were treated with DNAse I for 15 minutes at 37°C.

HeLa, HEK293 and NSC34 cells were cultured in Dulbecco's Modified Eagle's Media (Euroclone) supplemented with 10% or 5% fetal bovine serum, 200 U/ml penicillin, 200 μg/ml streptomycin and 2 mM glutamine. SH-SY-5Y cells were cultured in Minimum Essential Media (Euroclone) supplemented with 10% Fetalclone III (Hyclone), 200 U/ml penicillin, 200 μg/ml streptomycin and 2 mM glutamine. All cultures were grown as a monolayer in a humidified incubator at 37°C in an atmosphere of 5% CO₂.

Transient DNA transfections were performed by using Lipofectamine 2000 (Invitrogen) according to the manufacturer's specifications. Briefly, 8 × 10⁴ cells per well of a 24-well plate were seeded the day prior to transfection. Cells were transfected with DNA (500 ng) and cultured for an additional 24 or 48 hours. In cotransfection experiments, pRL-CMV DNA was added in 1:100 ratio.

mRNA transfections were performed by using Transmessenger Transfection Reagent (Qiagen) according to the manufacturer's specifications. Briefly, 8 × 10⁴ cells per well of a 24-well plate were seeded the day prior to transfection. Cells were transfected with mRNA (2 μg) and cultured for an additional 8 hours.

Immunofluorescences were performed as described previously 3.

Cells were scraped in phosphate-buffered saline and lysed for 30 minutes in RIPA buffer (50 mM Tris-HCl, 1% NP-40, 0.25% Na-deoxycholate, 150 mM NaCl, 1 mM EDTA, pH 7.4) and protease inhibitor cocktail (Sigma-Aldrich) in ice. Protein samples were resuspended in SDS sample buffer and subjected to standard sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) followed by protein transfer to a polyvinylidene difluoride membrane (Amersham). Spastin was revealed by immunoblotting with S51 polyclonal antibody 7.

Total RNA was extracted from cells using TRIzol (Invitrogen) according to the manufacturer's specifications. Rapid amplification of 5'-ends cDNA was carried out using a FirstChoice RLM-RACE kit (Ambion) according to the manufacturer's instructions with the following exceptions. The outer PCR reaction was carried out with 10 pmol gene-specific outer primer (5'-ACCATTCCACAGCTTGCTCCTTCT-3'), 1.25 units of Pfu Ultra (Stratagene) and 1.5 ng first-strand cDNA reaction. The PCR conditions were as follows: (1×) 94°C, 3 minutes; (35×) 94°C, 30 seconds; 55°C, 30 seconds; 72°C, 90 seconds; (1×) 72°C, 10 minutes. The inner PCR reaction was carried out with 10 pmol gene-specific inner primer (5'-CGCAAGCTTAGGCCTGTTTGTGGAAGACTCGGACG-3', BamHI site in bold), using the same conditions as for the outer PCR. PCR products were separated on 2% agarose gel, cloned BamHI into a pBluescript KS-vector (Stratagene) and sequenced.

Conservation studies were performed using the human BLAT search database 33. Alignments between human and mouse sequences were performed by using mVISTA 34 with the following parameters: min_Y (minimum Y value on the mVISTA plot) 40%, min_id (minimum conservation identity) 50%, min_length (minimum length for a CNS) 50 bp. IRES prediction was performed by using UTRScan 35. Transcription factors binding sites were predicted using two different matrixes. MATCH™ 36 parameters: profile, vertebrates; cut-off selection, minimize the sum of both error rates. PATCH™ 36 parameters: sites selection, vertebrate sites; minimum length of site, 10 bp; maximum number of mismatches, 0; mismatch penalty, 100; lower score boundary, 87.5. Bio-informatic analysis of 5'-end full-length SPG4 cDNas was performed using the CAGE analysis website 37 and the DBTSS 38.

Data are expressed as the mean ± standard error of the mean. Statistical analysis was performed using a two-way unpaired Student's t test.