Referee 2's review:

**********
N.B. These are the comments made by the referee when reviewing an earlier version of this paper. Prior to publication the manuscript has been revised in light of these comments and to address other editorial requirements.
**********

In this manuscript entitled “High-efficiency stem cell fusion-mediated assay reveals Sall4 as an enhancer of reprogramming”, the authors suggested that the efficiency of fusion mediated reprogramming could be improved by using different fusion methods. For example, the reprogramming of MEFs can be further improved by Sall4 over-expression in MEFs prior to fusion. The present study investigating the role of pluripotencyassociated genes (Oct4, Nanog, and Sox2) as well as the newly identified Sall4 gene via a cell fusion assay may be a good model system for enhancing our understanding of early reprogramming events during cell fusion.

While the experiments may yield interesting results, they do suffer from the technical flaws. For example, the screening of reprogrammed hybrid cells is not convincing, as the authors checked the reprogramming efficiency by FACS without taking into account the autofluorescent cells. Moreover, the GFP-positive cells showed in Figures 1B and 1C appear to be autofluorescent MEFs, as the GFP signal is only detected in cytoplasm but not in the nucleus, which is a typical autofluorescent artifact. For a GFP signal not to be considered a false positive, it must be detected in both the cytoplasm and nuclei of cells. The negative control wild-type MEFs fusion hybrid cells, which also displayed GFP positivity (Fig. 1E), provides corroborating evidence that cells that appear to be GFPpositive may not really be authentic GFP-positive cells. The GFP positivity displayed by wtMEFs may actually be due to contaminating germ cells (expressing Oct4) that were not removed during the MEF preparation. This potential source of error has not been addressed anywhere in the manuscript, and thus the experiments were erroneously reproduced throughout. For example, in Figures 4 and 5, the number of GFP-positive cells in the negative control is 1 and about 0.03%, respectively, but it should be 0 in both places. If it is not 0, this result indicates that the MEF cell population is not pure, but actually comprises a mixture of Oct4-GFP positive (pluripotent) and negative (nonpluripotent) cells, and as such cannot be used to screen reprogramming.

Due to this experimental flaw, the ensuing data cannot be considered valid.

Specific points:
1. The authors noted that cell fusion in adherent cultures increases the reprogramming efficiency by 15- to 590- fold compared with a previous protocol. However, no direct comparison of two fusion methods has been made in the current manuscript. Potential differences in the experimental conditions among studies preclude such comparisons from being made. The relevant paragraph should be removed from “Abstract”. In addition, the fusion method presented is similar to Helen Blau’s protocol. The authors should consider citing the paper by Helen Blau and thoroughly describe the fusion method used, as it is the major point of the paper.

2. The expression and localization of Oct4-GFP transgene appear to be restricted to the cytoplasm (Fig. 1C). However, in other figures showing activation of Oct4-GFP, the GFP signals in the hybrid colonies are evenly distributed throughout the cell; ie, in the both nucleus and cytoplasm. Thus, the authors should consider explaining the localized Oct4- GFP signal result. Otherwise, it would simply appear to be autofluorescence. An autofluorescent signal can easily be checked by observing the green cells through a red fluorescence filter. So, to avoid observing autofluorescence, the authors need to ensure that the cells are not detected through a red fluorescence filter. In Figure 1B, the cells shown at 24 and 48 h appear to be autofluorescent MEF cells. Therefore, although the cells contain 2 nuclei (24 h), the authors cannot determine whether they are hybrid cells or MEF-MEF fusion cells.

3. In Figure 1E, the authors show the FACS analysis of the reprogrammed cells. The authors mentioned that the left panel shows hybrids between ES and MEF (no Oct4-GFP) and right panel shows the reprogrammed hybrids between ES and MEF (with Oct4-GFP). However, there are also GFP-positive cells in the left panel. These MEFs do not contain the GFP transgene and should therefore be negative for GFP. The positive finding however clearly indicates that there is an experimental error. Moreover, in the text, the authors state that the percentage of GFP-positive cells was 0.029%, but this percentage appears to be that of the non-transgenic MEFs (see Fig. 1E, left panel). The authors also mentioned that they determined the percentage of GFP-positive cells at both 24 and 48 h, but only the FACS data is shown (Fig. 1E).

4. In Figure 2B, the authors showed the protein levels of individually infected genes by
Western blot, but the levels cannot be meaningful unless they are compared with those of a control gene (eg. beta-actin).

5. On page 7, 2nd paragraph, the authors state, “With G418 drug selection, GFP-positive colonies began to appear within 5 days post-fusion.” Why is reprogramming so delayed following G418 selection?

6. On page 8, line 1, the authors suggested that expression of Nanog, Sox2, and Sall4 leads to a significant increase in the number of reprogrammed colonies, but in Figure 3C, we see no significant differences between control mCherry (by error bars) and 3 separate samples. The authors have not indicated the statistical analysis used to determine this apparent “significant increase…”.

7. The authors also infected these viral constructs into ES cells, with no apparent effect on Oct4 activation. The silencing of viral construct (CMV promoter) already has been described by many other researchers. So the authors may use a cellular promoter to investigate the role of a gene of interest in ES cells (Fig 4D).

8. In Figure 5B, error bars between control mCherry and Sall4 appear to overlap, or not be significant.

Minor points:
1. The authors concluded that over-expression of Oct4 alone may inhibit reprogramming, as they did not observe an improvement in the reprogramming rate following Oct4 overexpression alone. However, this conclusion is not warranted, considering that infection of cells with Sox2 and Nanog individually also didn’t show any significant enhancement of reprogramming (Figs. 3A, 3B), although an increase in colony forming rates was observed. The following statement must be removed, “Oct4 over-expression inhibits the reprogramming of somatic genome”.

2. On page 9, line 9, the authors described that “fluorescent intensity of the negative control, mCherry, was significantly lower in infected ESCs than in infected MEFs from previous experiment;…”. How was the fluorescent intensity measured?

3. The fact that Oct4-gfp activation occurs within 24-48 h after fusion is not a novel finding. Previous studies have shown that Oct4-gfp can be activated with the first cell cycle within 24 h post-fusion (Han et al., Stem Cell Online published).

4. In Figure 4, the Y-axis represents “relative GFP expression”. Does this refer to the number of GFP-positive cells or to the intensity of GFP signal? This needs to be clarified.