ATP depletion plays a pivotal role in self‐incompatibility, revealing a link between cellular energy status, cytosolic acidification and actin remodelling in pollen tubes

Summary Self‐incompatibility (SI) involves specific interactions during pollination to reject incompatible (‘self’) pollen, preventing inbreeding in angiosperms. A key event observed in pollen undergoing the Papaver rhoeas SI response is the formation of punctate F‐actin foci. Pollen tube growth is heavily energy‐dependent, yet ATP levels in pollen tubes have not been directly measured during SI. Here we used transgenic Arabidopsis lines expressing the Papaver pollen S‐determinant to investigate a possible link between ATP levels, cytosolic pH ([pH]cyt) and alterations to the actin cytoskeleton. We identify for the first time that SI triggers a rapid and significant ATP depletion in pollen tubes. Artificial depletion of ATP triggered cytosolic acidification and formation of actin aggregates. We also identify in vivo, evidence for a threshold [pH]cyt of 5.8 for actin foci formation. Imaging revealed that SI stimulates acidic cytosolic patches adjacent to the plasma membrane. In conclusion, this study provides evidence that ATP depletion plays a pivotal role in SI upstream of programmed cell death and reveals a link between the cellular energy status, cytosolic acidification and alterations to the actin cytoskeleton in regulating Papaver SI in pollen tubes.


Fig. S1 Verification of [pH]cyt values from calibration of the pH indicator pHGFP in vivo using
propionic acid by comparison with those from a nigericin clamp-based calibration.

Table S1
Details of the transgenic Arabidopsis thaliana self-incompatible (SI)-lines used in this study.

Table S3
Quantification of cytosolic pH ([pH]cyt) and corresponding proton concentration in different distal areas of growing and self-incompatibility (SI)-induced Arabidopsis thaliana pollen tubes co-expressing PrpS1 and pHGFP.
Video S1 Time-lapse image series of a pollen tube from a 'rapid' Arabidopsis thaliana line coexpressing PrpS1 and Lifeact-mRuby2 after self-incompatibility (SI)-induction showing filamentous (F)-actin reorganization and foci formation.
Video S2 Time-lapse image series of a representative pollen tube from a 'rapid' Arabidopsis thaliana line after treatment with 2-deoxyglucose (2-DG) and antimycin A, showing timing of growth arrest, which is a key feature of self-incompatibility (SI).
Video S3 Time-lapse ratio-image series of a normally growing pollen tube from a 'rapid' Arabidopsis thaliana line co-expressing PrpS1, pHGFP and Lifeact-mRuby2 showing the distribution of cytosolic pH ([pH]cyt) during normal growth.
Video S4 Time-lapse ratio-image series of a pollen tube from a 'rapid' Arabidopsis thaliana line co-expressing PrpS1, pHGFP and Lifeact-mRuby2 after self-incompatibility (SI)-induction, showing spatiotemporal changes in cytosolic pH ([pH]cyt) during the SI response.  Wilkins et al. (2015) used a similar in vitro protocol, with slight changes, and also obtained a [pH]cyt of 5.5 after SI. In this paper we performed in vivo calibrations in pollen tubes, using propionic acid buffer to manipulate the [pH]cyt in pollen tubes expressing pHGFP.
SI-induced acidification of [pH]cyt to 5.5 is conserved regardless of reporter or calibration method.   characteristics were very similar. We therefore identified four key 'stages' as categories that were typical of the alterations observed. We numbered the different stages to allow comparison between events observed in the 'slow' lines, as there was considerable variation in timing of the response. Stage 1 looked essentially like untreated pollen tubes. At stage 2, long longitudinal F-actin bundles/cables were still apparent in the central region, but the cortical region had short F-actin filaments that were no longer aligned with the growth axis.

Fig. S2 Live-cell imaging shows key features of filamentous (F)-actin alterations triggered by self-incompatibility (SI) in
At stage 3, very few intact actin filament cables were observed. Stage 4 had large punctate actin foci. UT, untreated. Scale bar, 5 μm.

actin foci formation in pollen tubes from a 'rapid' Arabidopsis thaliana line. (a)
Quantification of F-actin foci that were formed following two different types of progression.
Type I: Foci formed from short randomly orientated F-actin fragments that aggregated relatively rapidly; Type II: Foci formed directly from severed actin filament cables, with fragments of actin bundles aggregating to form the distinctive large foci, or with small foci subsequently coalescing to form larger foci. Error bars indicate ± SD. More than 80 foci were categorised per pollen tube (over 12 pollen tubes undergoing the SI response were analysed in total).

Fig. S6 Timing of pollen tube growth arrest after ATP depletion. Plots showing pollen tube
growth and the timing of growth arrest, which is a key feature of self-incompatibility (SI). The growth of three representative pollen tubes is shown before and after ATP depletion using 10 μM antimycin A and 15 mM 2-deoxyglucose (2-DG). Arrows indicate the timepoints when treatment was added. As expected, the ATP depletion drugs led to almost instantaneous growth arrest, within 1 min.

Fig. S7 Pollen tubes do not exhibit elevated caspase-3-like/DEVDase activity during the selfincompatibility (SI)-or ATP-depletion induced acidification time-period. (a) Caspase-
3/DEVDase activity in pollen tube extracts from an Arabidopsis 'rapid' line (pntp303::pHGFP_pntp303::PrpS1) after SI induction. 100% activity represents zero (baseline) activity in unchallenged pollen tubes. No significant increase in DEVDase activity was observed 20 min after SI induction in Arabidopsis pollen tubes compared to normally growing control pollen tubes (with growth medium (GM) added), while after 2.5 hours of SI induction, a significant increase in DEVDase activity was detected (n = 3). Error bars indicate ± SD. This provides evidence that SI-induced pollen tubes are not dying during the acidification period.
(b) Caspase-3/DEVDase activity in Arabidopsis pollen tube extracts treated with ATP depletion drugs. 100% activity represents zero (baseline) activity in unchallenged pollen tubes. Five hours after addition of the ATP depletion drugs (2-deoxyglucose (2-DG) + antimycin A) to pollen tubes of an Arabidopsis 'rapid' line, way after the time frame of when the ATP depletion was observed, there was no evidence of caspase-3/DEVDase activity detected (n = 3). After 5 h treatment with these two drugs together, Arabidopsis pollen tubes had caspase-3/DEVDase activity that was not significantly different from control, normally growing pollen tubes that were treated with GM. Error bars indicate ± SD. This provides evidence that artificial ATP depletion using these drugs does not trigger caspase-3/DEVDase activity, even long after the time-frame when the ATP depletion was observed.

Table S1 Details of the transgenic Arabidopsis thaliana self-incompatible (SI)-lines used in this study. Two types of transgenic Arabidopsis lines with different levels of expression of
PrpS in the pollen (see Wang et al. 2020) were used in this study. The 'rapid' lines express a relatively high level of PrpS and exhibit a fast SI response. For example, most pollen tubes from a 'rapid' line exhibited punctate foci within 10 minutes of SI induction, while this would normally take one hour in Papaver pollen tubes using a similar concentration of PrsS recombinant protein. The 'rapid' line was therefore useful for visualizing SI-induced alterations continuously from start to finish in a single pollen tube, which would otherwise be impractical with normal expression levels of PrpS. The 'slow' line expresses relatively low level of PrpS and has a slower SI response. Most pollen tubes from this line exhibited punctate foci after 1-1.5 hours of SI induction. This was useful to study the shortening rate of actin filament cables induced by SI. The genetically encoded actin probe Lifeact-mRuby2 (Riedl et al., 2008;Bascom et al., 2018) and pHGFP (Moseyko and Feldman, 2001) were co-expressed in the 'rapid' line 3 to enable imaging of F-actin and [pH]cyt after SI.

Table S2 Propidium iodide (PI)-staining of Arabidopsis thaliana pollen tubes from the 'rapid'
line co-expressing PrpS1 and Lifeact-mRuby2 after self-incompatibility (SI) induction. None of the pollen tubes observed (n = 1338 total) had PI-stained nuclei at any time point up to 30 min after SI induction, which was when cellular [ATP]i and [pH]cyt had already decreased and plateaued. At 60 min after SI induction only ~10 % of pollen tubes were PI-positive, but by 3 h after SI-induction 96% (n = 265) of observed pollen tubes were PI-positive (n = 265 pollen tubes observed for 180 min). This demonstrates that no permeabilization of the plasma membrane was observed within the time-frame of the significant [pH]cyt decrease observed in these studies. This provides evidence that the SI-induced pollen tubes were not dying at this early stage of SI and membrane integrity was lost much later. These data provide good evidence that cytosolic acidification occurred upstream of SI-induced programmed cell death (PCD) rather than being merely a consequence of PCD. The pseudocolour shows pH values, with 'hot' colours indicating high pH and 'cool' colours indicating low pH. Note the peripheral acidic patches appearing at ~2-3 min and subsequent cytosolic acidification.