The manuscript by Denus and colleagues reports a sensitive assay to study unconventional secretion of cytosolic proteins lacking signal sequence, with an emphasis on misfolded proteins relevant to neurodegenerative diseases. The assay involves the fusion of a luciferase peptide to cargo of interest, and then express these proteins in a neuroblastoma cell line. The release of these fusion proteins to the medium can be detected using a split luciferase assay kit. They used this method to study the secretion of AD associated Tau mutants and ALS-associated SOD1 mutants. They also used the assay to perform a small scale screen, focusing on known AD risk factors. Overall, this study developed a useful tool for the field and also reveals some interesting connection between Abeta peptide and tau secretion. The paper experiments were carefully conducted and the data are convincing. The following issues should be addressed to improve the clarity of the presentation.

Main points:
They should explain more clearly how the IL1beta construct was designed. Where is the fusion part inserted? The diagram in Figure 1 shows that the SBP and HiBit were placed at the N-terminus of IL1beta, presumably downstream of the signal sequence. Since it is known that IL1beta needs to be cleaved by Caspase, resulting in a mature C-terminal domain, which is released. If SBP and HiBit were placed out of the mature domain, it should not be released by UcPS. Please clarify this point.
Since the authors claim that one advantage of the new assay is its quantitative power, it will be important to reveal the exact percent of these cargos released by SHS-Y5Y cells using proteins in lysate as a normalization.
On Page 7, the authors described the use of the detergent digitonin to measure the percentage of the cargos that were released via exosome or extracellular vesicles. The results seem to be much higher than previous reports (Wu, S. 2023 eLIFE, Lee, J, et al., 2016 NCB). The experiment described here needs an additional control, that is to show that the detergent did not improve the assembly efficiency between the HiBit and LgBit in the context of misfolded proteins like Tau, SNC and SOD1. They should use ultracentrifugation to obtain a pure cytosolic fraction and treat it with digitonin similarly. There is a possibility that the added fusion peptide HiBit may form inappropriate interaction with misfolded Tau, SOD1 or SNC, which reduces the assembly efficiency until digitonin disrupts these interactions.
In Figure 3, I wonder why SNC was not included in the RUSH assay.
On page 10, since the SOD1 G93A also increased LDH release into the medium, they should also discuss that the release of this mutant may be in part caused by cell lysis.
The description of the siRNA screen on page 12, 13 should be revised to improve clarity. The authors mentioned that knockdown of several genes modulates Tau UcPS and the list includes both positive and negative regulators. They should separate them, so readers can see the connections or lack of correlation between certain cellular pathways and Tau UcPS. For example, it is quite confusing to say that genes involved in endolysosomal pathways and APP processing strongly impact Tau UcPS, but it turns out that some genes in these pathways are positive regulators while others are negative. More specific information should be given to better explain these observations. I also noticed that Presenilin 1 appears to be a positive regulator while Presenilin 2 is a negative regulator, although these proteins are homologous to each other. There is no explanation whatsoever on this.
Is the Abeta peptide used in the last figure Abeta42? They showed that this peptide enters the endocytic system and it seems to influence Tau UcPS via this system. If so, how can the authors reconcile this finding with the studies showing Tau secretion by type I UcPS, as mentioned in the introduction.

The authors have addressed all my comments.