It has been reported that the commonly used method of extraction of photosynthetic pigments with aqueous acetone sometimes results in artifactual chlorophyllide formation [18, 23]. In the present study, we determined the quantity of chlorophyllide formation before, during or after extraction of these pigments using several different methods of extraction. By suppressing CLH activity during extraction, we demonstrated that only trace amounts of chlorophyllide, if any, are present in cells prior to extraction (see Figures 4, 5, 7 and 8). We also showed that both chlorophyll and chlorophyllide are stable in acetone after extraction (Table 1). Therefore, it is unlikely that chlorophyllide is formed in the solvents after the extraction procedure is completed. Based on our collective results, we concluded that chlorophyllide is formed during the extraction process. We speculate that chlorophyllide is formed when acetone infiltrates the tissue, or when the tissue is homogenized in acetone. During these processes, the actual concentration of acetone to which cells are exposed may increase gradually rather than immediately, thus allowing an opportunity for aberrant enzymatic reactions to occur. Although CLH is known to precipitate in pure acetone, it is capable of remaining highly active in lower concentrations of aqueous acetone [18, 27]. Therefore, it is likely that CLH catalyzes the formation of chlorophyllide during extraction until the actual acetone concentration reaches nearly 100%. This hypothesis explains the differential effects of DMF on chlorophyllide formation during chlorophyll extraction from different plant species (Figures 7 and 8). DMF suppressed chlorophyllide formation in Arabidopsis leaves almost perfectly, while it allowed chlorophyllide formation in other plant species whose leaves are thicker than Arabidopsis (Figure 8). These observations can be explained by the assumption that the infiltration of DMF occurs more slowly in thicker leaves as compared to in thinner leaves.
The aforementioned hypothesis raises the question why CLH is only active after the tissue is homogenized with organic solvents or soaked in organic solvents. A possible answer to this question may be that CLH is active in cells but separated from chlorophyll in cells. Schenk et al.  used CLH-GFP targeting experiments and confirmed that CLH is localized outside of chloroplasts. If CLH is indeed separated from chlorophyll in intact cells, the homogenization of cells or immersion of tissue in acetone may disrupt cell structures and enable CLH to act on chlorophyll.
Chlorophyllide has been long considered to be an intermediate of both chlorophyll biosynthesis and breakdown [20, 28]. Hörtensteiner and co-workers [23, 29], however, suggested that chlorophyllide is not a true intermediate of chlorophyll breakdown, at least during leaf senescence in Arabidopsis. Instead, they indicated that chlorophyll is degraded via pheophytin (Figure 1B). Our results are consistent with the chlorophyll breakdown model of Hörtensteiner and coworkers [23, 29]. The majority of chlorophyllide detected in acetone extracts of leaf pigments in our experiments were formed by the action of CLH during extraction (see Figures 3 and 4). These results suggest that plants only accumulate a small amount (if any) of chlorophyllide in cells under either normal growth conditions ( and this study) or when exposed to MeJA.
We compared three methods that suppress CLH activity with a conventional acetone-extraction method. In the first method, Arabidopsis leaves were boiled for a short time (5 or 10 sec). This procedure almost completely suppressed chlorophyllide formation with Arabidopsis and G. coronaria leaves. Bacon and Holden  already reported that a 5 minute period of boiling eliminates chlorophyllide formation. Their boiling time, however, appears to have been too long since they observed extensive decomposition of the pigments . In principle, the boiling time used in this procedure should be optimized for each plant species but we do not suggest boiling leaves for more than 10 sec for most plant species (see Figure 8). Thicker leaves may necessitate a longer boiling time. For example, we found that a 30 sec boiling time worked well to eliminate CLH activity in mulberry leaves in our laboratory (data not shown). This method appears to have another advantage in increasing the extraction efficiency of pigments from thicker leaves such as pea leaves when pigments are extracted by immersing leaves in organic solvents (see Figure 8C). Thus, the boiling method combined with the use of DMF as an extractant would be worth testing when pigments are extracted from thicker leaves. A possible drawback of the boiling method is the potential for additional types of modification to chlorophyll molecules. For instance, we observed a slight increase in pheophytin a concentration in our extracts (Figure 6) indicating that 0.1 to 0.2% chlorophyll a might be converted to pheophytin a by boiling. Thus, the boiling method is recommended in studies where the quantitation of pheophytin a is not being considered.
In the second method, frozen leaves were ground at sub-zero temperatures in a metal box that was cooled with liquid nitrogen. The leaves were then homogenized in pure acetone cooled to −30°C using an automatic bead shaker, Shake Master. The use of this shaker facilitates the processing of a relatively large number of samples. It is also possible to use cooled mortar and pestles for grinding leaves at sub-zero temperatures. However, this approach may be laborious and time-consuming when the analysis of a large number of samples is required. In addition, the recovery of a sufficient amount of solvent from a mortar can be problematic when only a small amount of sample tissue is used or available . Therefore, the usage of a mortar and pestle with this method is recommended only when a relatively small number of samples need to be analyzed and when a sufficient amount of tissue is available for each sample. Another limitation of this method will be a requirement of liquid nitrogen, which might not be readily available in field research. Regardless of these limitations, this method is superior to other methods in completely suppressing CLH activity in all plant species tested in this study. This method would be suitable for determining the minimum levels of chlorophyllide formation.
In the third method, pigments were extracted with DMF. This solvent has been previously used for pigment extraction [24, 26] but, to the best of our knowledge, was not tested for chlorophyllide formation. This solvent prevents CLH activity even during an overnight incubation of Arabidopsis leaves at 4°C (Figure 7). Therefore, the use of DMF appears to be the best option for extracting photosynthetic pigments from this model organism for downstream analysis using HPLC without introducing artifacts. However, this solvent is not as effective for G. coronaria, pea and cherry leaves as it is for Arabidopsis leaves (Figure 8). Moreover, this solvent is a possible liver toxin  and all appropriate safety guidelines should be adhered to in its use. Although the volatility of DMF is low, it should be carefully handled in an exterior venting fume hood. In conclusion, the use of DMF might be restricted to Arabidopsis or similar plant species under well-ventilated laboratory conditions.