Technical Comments

Response to Comment on “The Arabidopsis Circadian Clock Incorporates a cADPR-Based Feedback Loop”

See allHide authors and affiliations

Science  09 Oct 2009:
Vol. 326, Issue 5950, pp. 230
DOI: 10.1126/science.1169736


Xu et al. were unable to measure circadian oscillations of cyclic adenosine diphosphate ribose (cADPR). Their experiments showing very low concentrations of cADPR lack appropriate controls, which suggests that technical limitations might explain their negative result. Xu et al. also report that chemically induced ADP ribosyl cyclase did not alter clock function, which exactly replicates our findings.

We reported the existence of a feedback loop within the plant circadian clock that incorporates cyclic adenosine diphosphate ribose (cADPR) (1). Xu et al. (2) failed to report essential controls to demonstrate that they could detect the fold-changes in [cADPR] that we found to be caused by the circadian clock. For example, the plant hormone abscisic acid increases [cADPR] and acts as a physiologically relevant positive control (3). This is detected by the assay in our hands (Fig. 1). Xu et al. used chemical induction of heterologous ADPR cyclase expression (XVE:ADPRc) as a positive control, but this causes [cADPR] increases manyfold greater than endogenous alterations, so it is not an appropriate control to demonstrate assay sensitivity in the physiological range. Xu et al. reported this increased [cADPR] to 8 pmol mg−1 protein, lower than many estimates of [cADPR] in prestimulated cells (Table 1). The increase reported by Xu et al. after 24 hours of induction of XVE:ADPRc is 3- to 4-fold, compared to over 30-fold and 20-fold (1, 3) increases in the same lines at comparable time points in previous studies. Because the amplitude of circadian oscillations of [cADPR] are 5-fold smaller than those caused by XVE:ADPRc induction and Xu et al. could only measure small changes in response to XVE:ADPRc induction, we conclude that Xu et al. failed to measure alterations in [cADPR] with a sufficiently high signal-to-noise ratio to allow the possibility of measuring circadian [cADPR] oscillations.

Fig. 1

ABA increases [cADPR] in wild-type Arabidopsis. Three-week-old Col-0 Arabidopsis seedlings grown in 12 hours light/12 hours dark were treated with 50 μM ABA or 0.5% methanol control by flooding plates for 1 min. Three independent replicates were harvested at the times indicated. Each sample was measured at least twice. Points indicate the mean of the six measurements, with standard error bars shown.

Table 1

Prestimulated concentrations of cADPR reported from cell extracts vary by at least one order of magnitude. The concentrations reported by Xu et al. are at the lowest end of this range.

View this table:

In the absence of physiologically relevant controls, we can only speculate as to why Xu et al. failed to measure circadian oscillations of [cADPR] with the same assay that we used. cADPR could have degraded during extraction, but Xu et al. did not quantify cADPR recovery. They measured prestimulated cADPR concentrations (0.2 to 0.8 pmol mg−1 protein) that are the lowest values that we have found in any study using the units pmol mg protein−1 (the published range is 0.6 to 1500 pmol mg−1 protein) (Table 1). Possible explanations for very low [cADPR] reported by Xu et al. include cADPR degradation by contact with perchloric acid (4). Small differences in temperature and duration of perchloric acid exposure affect cADPR degradation rates profoundly during extraction (4).

Our reported experiments were designed to be accurate and unbiased. In our original report (1), figures 2 (A and B), 3 (A and B), and 4 (A to C, and E and F), as well as figures S2, S6 (A and B), and S8 (A and B) were from experiments designed to limit experimenter knowledge of the treatment. Extractions and assays for cADPR were performed on pseudo-randomized coded material. We reported two independent experiments with equivalent results that were representative of three experiments [figures 1A and 1B and figure S2 in (1)]. Our measurements of basal [cADPR] are within the range reported for plants and animals (Table 1). [cADPR] was measured by two researchers (A.N.D. and C.T.H.) in separate experiments that were consistent with each other and the literature. In a circadian-arrhythmic overexpressor of CIRCADIAN CLOCK ASSOCIATED1 (5), [cADPR] was arrhythmic and lower than basal Col-0 concentrations, which was consistent with [Ca2+]cyt dynamics (1, 6). We performed many controls to ensure that the assay functioned correctly. A plant-derived inhibitor did not interfere with the assay (Table 2). Recovery of cADPR standard after extraction was 71% and 85% at 5 nM and 500 nM, respectively. In extracts from 10 seedlings, measured cADPR was 232 nM with ADP ribosyl cyclase present in the assay and 7.61 nM without cyclase, indicating minimal contamination with nucleotides such as NAD+. Because our experiments limited experimenter bias and reported meaningful cADPR concentrations whose dynamics correlated with [Ca2+]cyt in both the wild type and CCA1-ox, we are confident that the circadian clock regulates [cADPR].

Table 2

Arabidopsis extracts do not inhibit the cycling assay for cADPR. The first three rows indicate the total concentration of cADPR that was measured from three separate extractions from five seedlings each. The fifth and sixth rows indicate the total assayed concentration of cADPR in extractions from five seedlings that were spiked with cADPR standard, at the beginning of the extraction procedure, to achieve concentrations of 200 nM and 20 nM in the extraction volume. If a plant-derived inhibitor was present, quantification of this internal standard would be suppressed.

View this table:

Xu et al. (2) claim that cADPR does not affect clock function because induction of Aplysia ADP ribosyl cyclase (ADPRc) did not affect circadian [Ca2+]cyt oscillations in chemically inducible Arabidopsis (XVE:ADPRc). This exactly replicates our findings [compare figure 4B in (1) with figure S6A in (2)]. Constitutive overexpression of 35S:ADPRc alters rhythms of leaf movement and [Ca2+]cyt [figure 4, C and D, in (1)], which is consistent with a previous report that cADPR regulates clock transcript abundance (3). Xu et al. claim their experiments with inducible lines demonstrate that our data obtained with 35S:ADPRc is irreproducible, but Xu et al. make a misleading comparison between inducible and constitutive manipulations because they have not performed experiments using constitutive lines. Whether chemical induction of expression pervades tissues comparably to 35S overexpression is uncertain. Future investigations of oscillator robustness to induced perturbations and stress signals will be extremely informative.

Xu et al. criticized our choice of concentrations of the cADPR signaling antagonist nicotinamide (10 to 50 mM). We demonstrated that these concentrations are required to inhibit ADPRc in Arabidopsis [figure S3 in (1)]. These are the concentrations used by previous researchers (7, 8) and the same range used by Xu et al. (30 mM) to reverse ADPRc activity in their cADPR assays. It would have been uninformative to use a lower concentration that does not inhibit cADPR synthesis when investigating signaling by cADPR. Because nicotinamide inhibits cADPR synthesis (9), and no other Ca2+ signaling antagonist tested affected the oscillations, a logical interpretation is that cADPR drives the circadian Ca2+ oscillation.

Xu et al. propose that nicotinamide alters clock function through known effects on poly(ADP ribose) polymerases (PARP), Sir2, or other ADP-ribosyl transferases. Although possible, this is not supported by published data. A study in mammals recently found that nicotinamide lengthens circadian period and that this occurs independently from the NAD+-dependent deacetylase SIRT1 (the mammalian homolog of yeast Sir2) (10). In animals, NAD+ oscillates with a circadian period (11, 12). The effect of nicotinamide also cannot be explained by inhibition of PARP. Nicotinamide inhibits PARP, decreasing [poly(ADP ribose)] (polyADPR) (13). This cannot explain the effects of nicotinamide on the clock because the Arabidopsis tej (PARG) mutant has high [polyADPR] and a long circadian period, and seedlings treated with the PARP inhibitor 3-aminobenzamidine have low [polyADPR] and a marginally shorter period (13), whereas nicotinamide-treated plants have low [polyADPR] but a long circadian period (1). Since nicotinamide affects NAD+ metabolism and cADPR is synthesized from NAD+, we cannot exclude other effects of nicotinamide on oscillator function because NAD+ modulates a variety of metabolic processes. Future investigations of plants with altered NAD+ metabolism, and the regulation of the oscillator by Ca2+, will determine whether nicotinamide exclusively modulates the circadian oscillator through cADPR or involves additional mechanisms.

Finally, a recent report demonstrated that in animals, pharmacological manipulation of suprachiasmatic nucleus ryanodine receptors (a cADPR-gated Ca2+ channel) alters circadian function (14) and that in Euglena ADPR cyclase activity oscillates with a day-night cycle, with maximal activity during the end of the day (15), similar to our finding that [cADPR] is highest in the subjective day in Arabidopsis. We provided several stand-alone lines of evidence that cADPR contributes to circadian [Ca2+]cyt rhythms and clock function in plants. It is reasonable to conclude that cADPR forms a feedback loop in the clock because (i) cADPR regulates circadian clock transcript abundance (3), (ii) the most statistically significant overlap between circadian- and signaling/hormone–regulated transcripts is with cADPR-regulated transcripts, (iii) [cADPR] is regulated by the clock, and (iv) high [cADPR] affects both [Ca2+]cyt and the clock. The data of Xu et al. (2) do not refute our findings that cADPR is regulated by and regulates the plant circadian clock.

  • * Present address: Department of Biology, University of York, York YO10 5YW, UK.

  • Present address: School of Biosciences, University of Exeter, Exeter EX4 4QD, UK.


View Abstract


Navigate This Article