Revisiting recent claims on the cellular expression, protein interaction, and functions of tomato cytosolic invertase2

Dear Editor,

We read with interest the publication of Zhang et al. (2024) entitled “Cytosolic invertase2 regulates flowering and reactive oxygen species-triggered programmed cell death in tomato.” There were several interesting findings described that we are intrigued by. However, we find that many of their experiments lacked essential controls with their claims unsubstantiated or misinterpreted. Hence, the authors’ conclusions are not supported. We outline several concerns below:

1. Clear evidence and controls are needed for claims of a neutral invertase located “exclusively in mitochondria.” The authors examined SlCIN2 protein subcellular localization by fusing it to mCherry and transiently overexpressed it in tobacco leaf epidermal cells. The authors stated that the fusion protein resided exclusively in the mitochondria (Fig. 1D). However, no positive localization control with a known mitochondrial marker was included, making it impossible to determine the subcellular localization. Moreover, the mCherry signal was extremely weak, and it is not possible to conclude that the faint puncta are mitochondria.

2. Yeast two-hybrid data for protein–protein interactions are inconsistent with separate compartments in vivo. Sucrose transporter (SUT) proteins contain 12 transmembrane domains and are localized in membranes (Barker et al. 2000; Lalonde et al. 2004; Baker et al. 2016). Zhang et al. (2024) used the GAL4 yeast two-hybrid assay, a transcriptional reporter system that detects protein interactions within the yeast nucleus, to test protein interactions for SlCIN2. Remarkably, they detected an interaction between SlCIN2, which they claim is mitochondrial localized, and SlSUT2, which resides on the plasma membrane. How can they interact while physically separated? Their finding that a 12-membrane-domain-containing protein can be localized to, and function within, the nucleus, as required for transcription of the GAL4-regulated genes, is most surprising and needs validation.

3. Subcellular markers are lacking for claimed localizations of SICIN2. The authors perform bimolecular fluorescence complementation assays in tobacco leaf epidermal cells to examine if SlCIN2 and SlSUT2 can physically interact. They claim that they do (Fig. 3D) but the images provided have exceptionally weak signals making it difficult to determine. The weak puncta that are visible raise questions about where in the cell this interaction would be occurring. The authors state that it is in the plasma membrane, but no bona fide markers (e.g. a known plasma membrane resident protein) were used to provide evidence for their conclusion.

4. Cell- and tissue-level expression specificity also lacks clear support. The authors examined the SlCIN2 cell-type specific expression in flowers using RNA in situ hybridization. They determined that the transcript is expressed in the anther phloem, and in epidermal cells and the phloem in petals (Fig. 1E). These conclusions are not well supported. In Fig. 1Ea, the expression was strong in all cells within the vein, with no obvious difference between phloem, xylem, or associated parenchyma cells. Perhaps in petals, there may be specific expression within vein cells (Fig. 1Ec), but their identity as phloem tissue, and which specific cell types they are, needs to be determined.

5. Focus is directed to one member of the SUT gene family but interpreted as if it were another. The authors performed RT-qPCR to examine gene expression for multiple SUTs in developing flower buds of a SlCIN2 overexpression line compared with wild type (Fig. 3A). They found SlSUT1 was upregulated ∼3×, whereas SlSUT2 and SlSUT4 were downregulated ∼3×. In the remainder of the paper, the authors focus exclusively on SlSUT2, surprisingly ignoring SlSUT1, which is known to control sucrose phloem loading in tomato (Hackel et al. 2006). SlSUT2 has been shown to reside on the plasma membrane (Reinders et al. 2002; Hackel et al. 2006) and is generally expressed in most tissues (Barker et al. 2000; Hackel et al. 2006). Additionally, antisense knockdown of SlSUT2 (previously called LeSUT2) caused pollen defects and prevented seed formation but did not affect sucrose phloem loading or cause any morphological defects (Hackel et al. 2006). Therefore, we think it is more probable that any changes observed in sucrose transport, plant growth, and floral development could be due to changes in sugar metabolism caused by SlCIN2 or sucrose transport by SlSUT1, and not by SlSUT2 as proposed by the authors.

6. Interpretation of work with the fluorescent sucrose analog would differ if the molecule could be metabolized by SICIN2. The authors examined sucrose transport activity of SlSUT2 in yeast using a fluorescent analog of sucrose, esculin (Reinders et al. 2008; Gora et al. 2012). No esculin uptake controls were included in the experiment (e.g. positive control using StSUT1 and negative control using an empty vector) making it difficult to evaluate the weak fluorescence shown (Fig. 3F). The authors also co-expressed SlSUT2 and SlCIN2 in yeast to examine esculin uptake. They measured an approximately 50% reduction in esculin (Fig. 3G) and concluded that SlCIN2 downregulates the sucrose uptake activity of SlSUT2. However, esculin is a structurally related compound to sucrose (Reinders et al. 2008; Gora et al. 2012), which is why it is able to be transported by some SUTs. We are unaware of any literature examining whether SlCIN2 may use esculin as a substrate; however, SlCIN2 binds and cleaves sucrose, so it is conceivable. Thus, an alternative explanation for the decrease in esculin fluorescence is that SlCIN2 has partial activity cleaving esculin, thus destroying its fluorescence. Appropriate controls are required to determine how SlCIN2 may impact the activity of SlSUT2.

7. A model is presented that misinterprets data and is inconsistent with known mechanisms of SUT functioning. In their model (Fig. 10), the authors proposed SlSUT2 functions to export sucrose from the companion cell to the apoplasm. However, SUTs function as sucrose–proton symporters (Boorer et al. 1996; Carpaneto et al. 2005). The pH of the apoplasm is approximately 2 units lower than the cytoplasm, indicating that there is 100-fold difference in proton concentration across the plasma membrane. Hence, SlSUT2 could not possibly function as a sucrose effluxer as proposed. Further, based on their in situ hybridization data for SlCIN2 (Fig. 1E), it is not possible to determine which vascular cells express SlCIN2, yet the authors proposed it is present in companion cells and an undefined anther cell, but not in parenchyma cells or sieve elements. No data are provided to support these speculations. Finally, the authors proposed SlCIN2 functioning in the tapetum cells although their in situ data do not support for this (Fig. 1E).

In summary, we find that many of the conclusions of Zhang et al. (2024) are premature and not supported by the data provided, especially regarding the claim that a mitochondrial localized SlCIN2 interacts with a plasma membrane localized SlSUT2 to regulate its activity.

Author contributions

Both authors drafted and edited the manuscript.

Funding

None declared.

Data availability

No new data were generated or analysed in support of this.