Rise in protease activity prior to visible senescence
A large increase in protease activity occurred in Iris tepals, starting from day 2 of vase life. The pH optimum of the in vitro reactions with the two tested substrates was consistent with the idea that most of the protease activity tested was located in an acidic compartment such as the vacuole. The effect of inhibitors, used in vitro, suggested that more than half of total protease activity was due to cysteine proteases. This mainly followed from the effect of E-64, which is purportedly specific. Inhibitors of serine proteases also reduced activity to c. 40%, indicating that serine proteases account for a considerable part of total activity. As inhibitors of metalloproteases only had a small effect, metalloproteases may play a minor role. The present results seem the first evidence for the presence of serine- and metalloprotease activity associated with senescence of petals.
Effects of proteasome and protease inhibitors on the time to visible senescence
A proteasome inhibitor delayed visible senescence by c. 12 h. Although this effect was small it is noteworthy, as it may indicate a role of the proteasome in the development of the senescence symptoms in tepals. Others have also found evidence for a role of proteasomes in senescence. During the differentiation of tracheary elements, proteasome inhibitors delayed programmed death during its commitment phase. These inhibitors had no effect during the lytic phase (Woffenden et al., 1998; Fukuda, 2000).
A few protease inhibitors tested delayed visible senescence: zinc chloride, aprotinin, AEBSF, and DFP. None of these inhibitors seems specific. Zinc ions inhibit cysteine proteases (Weis et al., 1995), but also some metalloproteases (Zuo & Woo, 1998), and a number of other enzymes such as nucleases (Gerhold et al., 1993). Zinc ions may also delay senescence by inhibiting homologues of caspases in animal cells (Enari et al., 1996). Aprotinin slightly delayed senescence and is known to inhibit serine proteases (Uchikoba et al., 1995), although it is not clear how specific this effect is. Both AEBSF and DFP are serine protease inhibitors and resulted in a considerable delay of visible senescence. DFP also inhibits deacylation reactions (Guther et al., 2001). In animal systems it inhibits esterases such as acetylcholinesterase (Testylier et al., 1999), and lecithin-cholesterol acyltransferase (LCAT), which hydrolyses oxidized polar phospholipids generated during lipoprotein oxidation (Goyal et al., 1997). DFP may therefore also have an effect through lipid metabolism. AEBSF also affects enzymes other than serine proteases. It slightly inhibited in vitro activity of papain, a cysteine protease (10–20% decrease; C. Pak, unpublished results). Furthermore, AEBSF has been reported to be an in vitro inhibitor of plant phospholipase D (Andrews et al., 2000). Its effect on the time to visible senescence (and on protease activity) may therefore be due to inhibition of lipid degradation. It seems that in Iris tepals large-scale lipid degradation precedes large-scale protein degradation (van Doorn et al., 2003). In neuron cells, AEBSF inhibited NADPH oxidase and attenuated the increase in oxygen free radicals (Hwang et al., 2002). In the same system, AEBSF delayed cell death by inhibiting caspase activity, but it remained unclear if this was due to inhibition at the caspase protein level (Rideout et al., 2001). It follows that both DFP and AEBSF can interact with several enzymes. Hence their mode of action in Iris tepal senescence, although suggesting inhibition of serine proteases is at present not clear. Whatever enzyme is inhibited by AEBSF, the reaction apparently depends on the presence of two contiguous CH2 residues between the amino residue and the benzene ring. Replacement of the two CH2 residues with one such residue and an extra lateral amino residue, in pABSF, resulted in loss of action (Table 1). When the activity of these two compounds on NADP oxidase was tested, both had a large inhibitory effect (Diatchuk et al., 1997).
The results may indicate that serine proteases are not only involved in bulk endoprotease degradation, but also are part of the processes that result in the large increase in total endoprotease activity. Following treatment with AEBSF (5 mm) and DFP (10 mm), starting on day 0, no rise in total protease activity was observed, at least until day 6. More than half of total protease activity on day 4 is apparently due to cysteine proteases, but AEBSF had only a small effect, in vitro, on cysteine proteases (C. Pak, unpublished; see this section). This means that AEBSF apparently inhibits the onset of the rise in total endoprotease activity. This is by contrast with application of AEBSF on day 2 or 3, which resulted in inhibition of only c. 50% of total protease activity. AEBSF may then only inhibit the serine protease proteins that have already been formed.
Although the data suggest that AEBSF regulates the onset of the increase in both serine- and cysteine proteases, it is unclear how it may do so. Very little is known about the regulation of the increase in total endoprotease activity. The effect of AEBSF may involve a regulatory serine protease. Apoptosis in several animal systems is regulated by granzymes (which are serine-proteases). However, we found that the granzyme inhibitors Z-AAD-cmk and DIC had no effect on the time to visible senescence in Iris. This we do not take as evidence against the idea that a serine protease – or even a granzyme homologue – is involved in the large increase in total endoprotease activity. We do not know how much of the granzyme inhibitors reached their potential targets, nor do we know how specific these inhibitors are.
The effect of AEBSF was by contrast to that of E-64d. If E-64d was included in the water on day 0 it inhibited total protease activity by c. 30%, which may coincide with direct inhibition of most cysteine proteases. E-64d therefore did not seem to regulate an early step in the onset of increased endoprotease activity.
The aim of the present work was to test two hypotheses: first that all four groups of endoproteases are involved in senescence-associated bulk protein degradation; and second that protease inhibitors delay the visible senescence symptoms. The first hypothesis was not fully substantiated as no positive evidence was found for aspartic protease activity. This conclusion is mitigated by the fact that we tested only one such inhibitor (pepstatin; other general aspartic protease inhibitors are apparently not available), but is reinforced by the finding that pepstatin is highly specific for aspartic proteases (Roberts & Taylor, 2003). The second hypothesis was substantiated, but only clearly so for purported serine protease inhibitors.
It is concluded that visible senescence of Iris tepals is preceded by a sharp increase in endoprotease activity. The onset of this increase showed a positive correlation with the onset of the visible senescence symptoms. In vivo data indicate that cysteine- and serine proteases, and probably metalloproteases, are involved close to the maximum of endoprotease activity. A number of purported protease inhibitors substantially reduced the increase in endoprotease activity and delayed the time to visible senescence. The timing of the rise in endoprotease activity therefore was correlated with the timing of the visible senescence symptoms.