New functions for electrical signals in plants

Electrical signals in plants have been known for over 100 years, but scientists are still looking for a function. The bulk of the work has focused on heat-wound evoked responses in above-ground parts, but evidence exists for their role in many processes, including transcription, translation and respiration. An additional function elegantly demonstrated in this issue of New Phytologist (Koziolek et al. pp. 713–720) is a transient reduction in photosynthesis. Koziolek et al. measured membrane potential, chlorophyll fluorescence and PSII quantum yield, and showed that the electrical signal preceded the changes in photosynthetic parameters – transport of chemicals in the phloem was far too slow to account for these changes. ‘One of the beauties of this study is that the authors recognize the need to take great care to avoid generating a wound during the act of measurement, and so use techniques imposing little or no stress on the plant’ For many years, the idea of electrical signaling in plants hovered in the research background, a minor interest topic even though Burdon-Sanderson (1873) and Charles Darwin (1875) had demonstrated the existence of electrical signals in insectivorous (‘motorized’) plants. However, by the early 1970s, with the publication of ‘The secret life of plants’ by Tompkins & Bird (1973) the credibility of the area had become severely strained in the research community making it essentially untouchable in the eyes and minds of funding agencies (Box 1). Just over 20 yr ago we presented some research showing that wounding one location on a plant caused very rapid changes in gene expression (translation/polyribosomes) in distant regions (Davies & Schuster, 1981). Joe Varner, a very eminent plant biochemist and molecular biologist, discussed this research with me and suggested that I might be looking at electrical signals. I remember my response: ‘But plants don’t have electrical signals, do they?’. Joe assured me that they did, and arranged for me to work in Barbara Pickard's lab at Washington University in St. Louis. I went there, did my first experiment by taping a plant in place on a styrofoam block, attaching electrodes, hooking them up to recorders, and then cutting the plant. The recorders, in sequence, starting with the one nearest the cut, exhibited a clear transient 70 mV change in membrane potential. I was now a ‘believer’ that electrical signals existed in normal plants, but was unsure whether they actually evoked any important responses. In order to ascertain whether there were any likely functions of electrical signals in plants, I spent a lot of time reviewing the literature until I was convinced that they really did exist, were worth working on, and might, in fact, explain several then inexplicable phenomena (Davies, 1987). Then came the breakthrough – the paper by Wildon et al. (1992) showing that electrical signals are involved in turning on protease inhibitor (PIN) genes in tomato. They showed that a local flame wound evoked rapid changes in membrane potential in distant tissue, while transport of chemical signals in the phloem took an hour or more. The electrical signals arrived in responding tissue much sooner than transcript accumulation began, while transcript accumulation preceded arrival of the chemical signal. The authors (Wildon et al., 1992) claimed, however, that the signal they were monitoring was a ‘genuine’ electrical signal, or action potential (AP). However, we had already worked on tomatoes and knew that a flame wound invariably evoked another type of ‘electrical’ signal known as a variation potential (VP). The AP is described as a ‘genuine’ electrical signal. It can be evoked electrically, it has an all-or-nothing character (below a threshold stimulus there is no response, while above this threshold, there is a maximal response), and it travels with constant velocity and magnitude (Zawadzki et al., 1991). The AP is a self-perpetuating signal based on the activity of voltage-gated channels which respond to (and cause) changes in membrane potential (Davies, 1993). By contrast, the VP cannot be evoked electrically, it varies with the degree of stimulus, and appears to ‘travel’ with decreasing magnitude and velocity away from the site of stimulus. The VP is a nonself-perpetuating signal based on mechano-sensitive (stretch-activated) and/or ligand-activated channels (Davies, 1993; Stankovic et al., 1997). Thus the VP is a local change (in living cells with ion channels) to either a hydraulic surge or chemicals transmitted in the dead xylem. Nevertheless, despite the misunderstandings concerning the different types of electrical signals in plants, this report (Wildon et al., 1992) helped make the study of electrical signals more respectable and closer to mainstream biology. Why do plants have electrical signals? Not only can this involve ‘What plant functions are governed by electrical signals?’, it also concerns ‘Why is there a need for these signals?’ The latter question can also be posed as ‘What properties do electrical signals have that chemical signals do not have?’ The first is ‘rapidity’. The VP in living cells is an almost direct aftermath of a virtually instantaneous loss of tension in the xylem, while the AP is a very rapidly transmitted signal predominantly in the phloem. Why might rapidity be important? In those instances where plants have a movement response, such as in the capture of insects, then speed is essential otherwise the prey would escape. Similarly, when a plant is attacked by insects (especially if they harbor a virus), the plant needs to put up systemic defenses as quickly as possible to prevent spread of the pathogen/herbivore. The second is ‘ubiquity’, insofar as all cells near the xylem will be affected by the VP, and those near the phloem by the AP. Similarly, ubiquity is important in sofar as the entire plant needs to be defended against potentially hostile threats, there may be little point in leaving some regions susceptible while defending others. This underlies the systemic nature of electrical signals as well as the local responses to them. The third property is ‘information’. This is important because the plant may or may not actually ‘recognize’ the exact insult. This is possible for the flame wound, which is the ‘insult’ used in this paper and the one preferred by us to evoke predictable, large, rapid, electrical signals and responses at the level of gene expression. It is conceivable that the plant does, indeed, ‘recognize’ the flame wound for what it is – fire – a very damaging threat which is pervasive throughout many ecosystems. However, it is equally conceivable that the plant does not recognize specifically that the insult is a local flame and so mounts a general stress response. It should be noted here that there is a huge difference between electrical signaling in the animal nervous system compared with that in a plant. In the animal nervous system, the whole point of an AP is to send information from one specific location to another, without affecting the intervening tissue – rather like a telephone system (Davies, 1993). By contrast, both the AP and VP in a plant are designed to inform as much of the plant as quickly as possible so that all the intervening tissue is informed – rather like a megaphone message warning everyone within hearing distance (Davies, 1993). The final property is ‘transience’. Not only is the signal (change in membrane potential) short-lived, but so is the response. After the wounded leaf has informed the rest of the plant, the recipient regions seem to ‘deal with the emergency and then get back to normal business’. Thus transience is important if the signal is merely to warn distant regions of impending disaster, and after receipt of such a warning, the (forewarned) distant tissue regains homeostasis. When studying wounding, great care needs to be taken to avoid generating a wound during the act of measurement, and so noninvasive techniques have to be used. One of the beauties of the study in this issue by Koziolek et al. is that the authors recognize this need, and so used techniques imposing little or no stress on the plant. They used standard techniques including microelectrodes (which impose a very slight puncturing stress) to measure electrical signals, a porometer to measure gas exchange, and autoradiography to estimate time of arrival of an chemical signal. The major feature, however, was their use of a recently developed method involving fluorescence imaging to measure the quantum yield efficiency of photosystem II. These fluorescence images are very persuasive. They show that PSII efficiency is transiently inhibited in the distant leaf beginning within 90 s of the heat pulse, traveling from the base to apex of the leaf (pinna) before returning to normal after about 450 s. The images are magnified to show that the vein region is inhibited initially, implying that the vein is the pathway of transmission, but that the intervein region is affected more massively, implying greater photosynthetic activity in that region. They show that generation and transmission of the electrical signal (variation potential) precedes the changes in stomatal conductance and gas exchange in fixed leaflets, which in turn slightly precedes inhibition of PSII quantum yield. Furthermore, the fact that wound responses tend to be transient adds another facet to the experimental approach. If the techniques require tissue extraction or fixation (as ours mostly do), then very detailed time courses need to be made – which adds to the time and expense of obtaining statistically reliable data. This problem can be circumvented by using ‘real-time’ imaging techniques, where the sample can be assayed repeatedly or even continuously. The authors achieve this, not only with their use of microelectrodes and porometers, both of which give continuous recordings, but especially with chlorophyll fluorescence measurements, which gives beautiful (false)-color images. In this context, it is worth noting that a new, non-invasive technique might have become available to monitor electrical signals. Recently, Hanstein & Felle (2004) described a ‘nanoinfusion’ technique where microelectodes are immersed in liquid infused into the substomatal cavity to measure membrane potential – in a totally non-invasive way – thus circumventing the (minor) damage elicited by micoelectrodes inserted into cells. The bulk of research in the area has focused on responses in above-ground parts and evidence exists for their role in many processes, including transcription (Wildon et al., 1992), translation (Stankovic & Davies, 1997) and respiration (Filek & Koscielniak, 1997), and in all cases the responses are transient. This paper (Koziolek et al.) adds inhibition of photosynthesis to this list, and also shows that PSII efficiency, as well as changes in gas exchange and stomatal aperture, are transient. It is still not known whether the transience of the response is a direct result of the transience of the signal itself (the change in membrane potential and concomitant fluxes of ions) or whether it signifies that the plant has successfully ‘coped with the problem’, or whether both of these are true. Nevertheless, there do seem to be some underlying similarities between the results from several labs working on many different responses to flame- (or heat-pulse-) induced variation potentials. The values shown here for lag times, time of maximum response, and time of return to baseline were roughly 90 s, 240 s, 450 s for PSII inhibition, gas exchange and stomatal conductance. These values are quite similar to, although somewhat slower than, the corresponding times for enhancement of respiration (Filek & Koscielniak, 1997) of 20 s, 80 s, 360 (s). We were also impressed by similarities in the kinetics of response of inhibition of photosynthesis shown in the paper with those from our own work on flame-wound induced changes in gene expression in tomatoes. Here we have measured the level of a putative second messenger, inositol phosphate (IP3) as well as accumulation of the transcript encoding the Rubisco small subunit. As shown in Fig. 1, the systemic increase in IP3 levels shows a lag of 75 s, a maximum at 90 s, and a return to the baseline by 120 s, while the Rubisco SS transcript shows a lag of 90 s, a maximum at 120 s and a return to the baseline by 150 s. There are two main differences between these results on transcript accumulation (Fig. 1) and those on respiration (Filek & Koscielniak, 1997) and photosynthesis (Koziolek et al.). First, our system has the disadvantage that we cannot assay the same tissue over time, but must harvest tissue for each time point. Second, we have an advantage insofar as we have identified a putative local signal (IP3) that could have been evoked by the systemic variation potential, and which could, in turn, evoke transcript accumulation. Three-wk old tomato plants were flame-wounded on leaf 3 and leaf 4, harvested at 15 s intervals, frozen immediately in liquid Nitrogen, and finely pulverized. One aliquot of the frozen powder was extracted to assay IP3 using the release of labeled IP3 from rat brain IP3 receptors. The other aliquot was extracted, electrophoresed, blotted and probed with a Rubisco small subunit probe according to published methods (Vian et al., 1999). Data are from three independent experiments, where the standard deviation was < 12% at all time points (Salinas-Mondragon, Atkinson and Davies, unpublished).