CERs reached zero at day 25 in the high CO 2 –grown leaves versus day 35 in the ambient-grown leaves.
These results suggest that the decline in photosynthetic rates characteristic of senescence is initiated at
an earlier time point in leaves that have an increased source strength; this is consistent with the feed-
back inhibition hypothesis. Interestingly, the onset of senescence in the high CO 2 –grown leaves oc-
curred while the leaves were still expanding.
An examination of other photosynthetic parameters provided further support for the notion that pho-
tosynthetic rates attain an earlier photosynthetic maximum in the elevated CO 2 –grown leaves [28]. These
included measurements of chlorophyll concentrations, Rubisco contents, and Rubisco activities. On the
basis of these results, we proposed a “temporal shift model” to explain the phenomenon of “acclimation”
(down-regulation of photosynthesis) that is frequently observed during growth of plants in elevated CO 2
[17]. In this model, the lower photosynthetic rates are the result of a shift in timing of the normal photo-
synthetic stages of leaf ontogeny to an earlier onset of senescence. Hence, when fully expanded leaves
from ambient- versus high CO 2 –grown plants are compared at a given day after leaf initiation (as in a typ-
ical “acclimation” experiment), lower photosynthetic rates are observed in the high CO 2 –grown leaves
because they are further along the progression of the senescence phase of development. Although there
appear to be species-specific differences, the findings of Miller et al. [28] are in general agreement with
other studies that have examined the impact of elevated CO 2 on leaf development [36–39].
If source strength has a regulatory role during leaf development, as suggested by the preceding stud-
ies, then it might be anticipated that a decreased source strength condition would have the opposite effect
and delay the initiation of the senescence decline in photosynthesis. To address this question, we exam-
ined leaf development in Rubisco antisense mutants of tobacco [13,27,40,41]. These plants have a de-
creased source strength because of a specific reduction in Rubisco content.
III. RUBISCO ANTISENSE MUTANTS
The Rubisco holoenzyme is composed of eight large subunit (LS) proteins coded for by single genes
(rbcL) on each of the polyploid chloroplast DNAs and eight small subunit (SS) proteins coded for by a
small multigene (rbcS) family in the nuclear DNA. To determine whether rbcL expression is responsive
to SS protein concentrations, as suggested by the “cytoplasmic control principle” [42], we used antisense
rbcS RNA to down-regulate the expression of rbcS messenger RNAs (mRNAs) and proteins in tobacco
[43]. For these experiments, tobacco plants were transformed with a highly expressed member of the to-
baccorbcS gene family cloned in reverse (antisense) orientation behind the cauliflower mosaic virus
(CaMV) 35S promoter. The resulting transgenic plants had reduced rbcS mRNA and SS protein levels.
The reductions in SS protein in these plants were matched by corresponding reductions in the accumula-
tion of LS protein and Rubisco holoenzyme. This lack of overproduction of the LS indicated that there
are stoichiometric alterations in the accumulation of the SS and LS in the mutant plants. In contrast to the
decreases in LS protein, rbcL mRNA levels were unperturbed in the mutants. This indicates that LS pro-
tein amounts are regulated posttranscriptionally in these plants. The various transgenic plants had a range
of Rubisco concentrations from 10 to 90% of normal, and the antisense rbcS RNA gene dosage correlated
inversely with Rubisco content.
To examine the nature of the posttranscriptional defect in LS accumulation, mutant plants were pulse
labeled with^35 S-Met [43,44]. LS synthesis was markedly decreased during the pulse, suggesting that the
antisense plants have a defect in rbcL mRNA translation. To pinpoint this defect, we examined polysome
profiles of rbcL mRNAs [44]. We found that rbcL mRNAs are associated with fewer than normal
polysomes in the antisense plants, suggesting that less LS accumulates because there is an impairment in
the initiation step of rbcL mRNA translation. This impairment appears to be specific for rbcL mRNAs
and not a general consequence of decreased plastid protein synthesis, inasmuch as the polysome distri-
butions (and abundances) of other plastid mRNAs are not affected in the mutants.
Our current working hypothesis is illustrated in Figure 2. In this figure, the SS (directly or indirectly)
affects the recruitment of ribosomes to rbcL mRNAs. For example, the SS could act as a translational ac-
tivator: increased SS would increase rbcL mRNA translation initiation and thereby increase LS protein
production (positive regulation). Alternatively, the LS (or its degradation products) could repress rbcL
mRNA translation initiation when the LS is produced in excess of the SS (negative regulation). This
mechanism would be similar to end-product inhibition at the translational level, as observed for some bac-
SOURCE STRENGTH AND LEAF DEVELOPMENT 119