The shady synphilin-1A was exposed when Simone Engelender and colleagues at Technion-Israel Institute of Technology were sifting through the expressed sequence tag (EST) databases at the National Center for Biotechnology Information website. EST sequences are usually deposited by scientists who have analyzed complementary DNA or messenger RNA, and as such, ESTs reflect RNA sequences that are transcribed, spliced, and presumably functionally important. But first authors Allon Eyal and colleagues noticed something odd about several synphilin ESTs. Three of them contained a 71-base sequence predicted to lie in an intron. In other words, that sequence should never end up in mRNA—in theory. In practice, however, it turns out that between exons 9 and 10 of synphilin lies another small exon that everyone else had missed. Eyal and colleagues coined it exon 9A.

But the mystery does not stop there. When Eyal and colleagues made an antibody to the peptide encoded by exon 9A, then used it to detect synphilin-1A in the brain, they found that the protein that bound to the antibody was only 75 kDa, about 25 kDa smaller than the predicted sequence.

Realizing that the 75 kDa protein is consistent with the loss of exons 3 and 4 through alternative splicing, the authors faced another enigma. Splicing out exons 3 and 4 would cause a frame shift that introduces a stop codon in exon 5, leading to a protein of less than 10 kDa. So how could synphilin-1A turn out to be 75 kDa? The answer lies in an apparently unique type of frame shift. Synphilin-1A uses a different start codon to synphilin-1, giving the A isoform a different N-terminus. But the splicing out of exons 3 and 4 resets the reading frame so that the C-terminal end of the two synphilin-1s are the same. “To our knowledge, two totally functional alternatively spliced transcripts originating from the same gene have never before been translated by the use of two different initial reading frames,” write the authors. If this turns out to be a more general phenomenon, then the number of proteins encoded by the genome could be markedly increased, the authors suggest.

Having figured out the size and nature of synphilin-1A, the authors then went on to characterize it more fully. They found that it is widely expressed in the brain, including in the substantia nigra, the region most affected in PD, and in the cerebral cortex, where synuclein inclusions are also found. Quantitatively, synphilin-1A represents slightly less than 15 percent of total synphilin-1. The A isoform is also found in synaptic vesicles and colocalizes with synaptophysin, suggesting that even with an alternative N-terminus, the protein may play an important role in the synapse.

But the protein may also be detrimental to neurons. When the authors overexpressed it in primary cortical cultures, they found that the cells retracted their processes. Markers of apoptosis also indicated that about 20 percent of the neurons had entered a cell death pathway. A smaller number of neurons also formed large aggregates, which, oddly enough, do not form when synphilin-1 is overexpressed in neurons, suggesting that the A isoform has a greater propensity to aggregate. The authors suggest that the shorter N-terminus on the 1A form might expose ankyrin-like domains on the protein, which could then interact with each other.

The researchers also observed an inverse relationship between the formation of inclusions and cell death. When they challenged neurons overexpressing synphilin-1A with MG132, the proteasomal inhibitor, about 40 percent of neurons formed organized inclusions. MG132 also increased cell death, but only in those neurons that had no inclusions. This finding speaks to a raging debate about whether inclusions are protective or detrimental. “Our findings clearly suggest a cytoprotective role for inclusion bodies,” the researchers write. Indeed, recent evidence suggests that inclusions may protect neurons expressing mutant huntingtin, the protein that causes Huntington disease (see ARF related news story).

Whether synphilin-1A plays any role in PD or other α-synucleinopathies has yet to be determined, but Eyal and colleagues did find α-synuclein in the synphilin-1A inclusions. They also found that the A isoform traps synphilin-1 in inclusions and prevents its degradation through the ubiquitin proteasome pathway. The finding suggests that even though it is not expressed as much as synphilin-1, the A isoform might pack a double punch, by trapping an equal amount of its more abundant twin, assuming a 1:1 binding.

As for patient data, Eyal and colleagues report that insoluble synphilin-1A is present in brain samples from DLBD patients, and it is also present in Lewy bodies from both PD and DLBD patients. Curiously though, in one out of three DLBD patient samples, the authors found that detergent-soluble synphilin-1A was totally missing, but as they point out, because of small sample numbers, no correlation can be made between soluble synphilin-1A and disease at this time.—Tom Fagan.

Reference:
Eyal A, Szargel R, Avraham E, Liani E, Haskin Y, Rott R, Engelender S. Synphilin-1A: An aggregation-prone isoforms of synphilin-1 that causes neuronal death and is present in aggregates from alpha-synucleinopathy patients. PNAS early edition. March 19, 2006. Abstract