SCARECROW gene function is required for photosynthetic development in maize

Abstract C4 photosynthesis in grasses relies on a specialized leaf anatomy. In maize, this “Kranz” leaf anatomy is patterned in part by the duplicated SCARECROW (SCR) genes ZmSCR1 and ZmSCR1h. Here we show that in addition to patterning defects, chlorophyll content and levels of transcripts encoding Golden2‐like regulators of chloroplast development are significantly lower in Zmscr1; Zmscr1h mutants than in wild‐type. These perturbations are not associated with changes in chloroplast number, size, or ultrastructure. However, the maximum rates of carboxylation by ribulose bisphosphate carboxylase/oxygenase (RuBisCO, V cmax) and phosphoenolpyruvate carboxylase (PEPC, V pmax) are both reduced, leading to perturbed plant growth. The CO2 compensation point and 13C‰ of Zmscr1;Zmscr1h plants are both normal, indicating that a canonical C4 cycle is operating, albeit at reduced overall capacity. Taken together, our results reveal that the maize SCR genes, either directly or indirectly, play a role in photosynthetic development. Significance statement SCARECROW (SCR) is one of the best studied plant developmental regulators, however, its role in downstream plant physiology is less well‐understood. Here, we have demonstrated that SCR is required to establish and/or maintain photosynthetic capacity in maize leaves.


| INTRODUC TI ON
The C 4 photosynthetic pathway is nearly always underpinned by characteristic Kranz anatomy, whereby vascular bundles are closely spaced and are often separated by only two mesophyll (M) cells (reviewed in Sedelnikova, Hughes, & Langdale, 2018). This cellular arrangement enables function of the C 4 cycle, in which CO 2 is initially fixed in the outer M cells by phosphoenolpyruvate carboxylase (PEPC) into the 4-carbon compound oxaloacetate.
Although there are at least three distinct C 4 enzymatic cycles (Furbank, 2011), all involve a 4-carbon derivative of oxaloacetate being shuttled to the inner bundle-sheath (BS) cells to be decarboxylated. The CO 2 released in the BS cells is refixed by ribulose bisphosphate carboxylase/oxygenase (RuBisCO) in the Calvin-Benson (C 3 ) cycle. Importantly, refixation occurs in a cellular environment where local levels of CO 2 are high enough to suppress the competing oxygenation reaction and thus to prevent the energetically wasteful process of photorespiration. As compared to ancestral C 3 species, the carbon concentrating mechanism that operates in C 4 species can lead to higher crop yields, as well as to improvements in water and nitrogen use efficiency (Brown, 1999;Sheehy et al., 2007).
Developmental regulators of Kranz anatomy have proved elusive; however, it was recently reported that the duplicated maize SCARECROW (SCR) genes ZmSCR1 and ZmSCR1h redundantly regulate Kranz patterning (Hughes, Sedelnikova, Wu, Becraft, & Langdale, 2019). In Zmscr1;Zmscr1h mutants, the majority of vascular bundles are separated by only one rather than two mesophyll cells, and there are additional patterning perturbations in both leaves and roots (Hughes et al., 2019). Roles for SCR genes in patterning processes are well established, particularly in Arabidopsis thaliana (hereafter referred to as Arabidopsis), in which AtSCR (the single copy ortholog of ZmSCR1 and ZmSCR1h) regulates the development of both the endodermis in the root and the BS in the leaf (Cui, Kong, Liu, & Hao, 2014;Di Laurenzio et al., 1996;Wysocka-Diller, Helariutta, Fukaki, Malamy, & Benfey, 2000). In addition to patterning defects, both Atscr and Zmscr1;Zmscr1h mutants are smaller than wild-type (Dhondt et al., 2010;Hughes et al., 2019), with reduced plant size In Atscr mutants associated with reduced cell proliferation rather than impaired root function (Dhondt et al., 2010). The effect of SCR on other downstream developmental and physiological processes, such as chloroplast development, photosynthesis, and growth, is less well understood.
To further investigate the role of SCR in shoot growth and development, we undertook a physiological characterization of Zmscr1;Zmscr1h mutant leaves. We show here that chlorophyll levels are reduced in Zmscr1;Zmscr1h double mutants, but not in either of the corresponding single mutants. Although transcript levels of the chloroplast developmental regulators ZmGLK1 and ZmG2 are reduced in Zmscr1;Zmscr1h leaves, there are no apparent perturbations to chloroplast development. The CO 2 compensation point and 13 C ‰ of Zmscr1;Zmscr1h leaves are also normal, suggesting that the C 4 photosynthetic cycle is operational, albeit with reduced PEPC fixation capacity and reduced maximum photosynthetic rate.
Collectively, these results imply a role for ZmSCR1 and ZmSCR1h (either directly or indirectly) in the establishment and/or maintenance of photosynthetic capacity in the maize leaf.

| Plant material and growth conditions
Both Zmscr1-m2;Zmscr1h-m1 (m2m1) and Zmscr1-m2;Zmscr1h-m2 (m2m2) double mutants were generated and described previously (Hughes et al., 2019). As segregating wild-type lines for both were indistinguishable, the WT (m2m1) line was chosen for comparison of both lines in this study. Plants were germinated and grown in a greenhouse in Oxford, UK with a 16 hr/8 hr light/dark cycle, with supplemental light provided when natural light dropped below 120 μmol photon m −2 s −1 . Day temperature was 28°C and night temperature was 20°C. Seeds were germinated in vermiculite and transferred after 9-11 days to 12 cm pots containing 3:1 mixture of John Innes No. 3 Compost (J. Arthur Bower) and medium vermiculite (Sinclair Pro). After 24 days, plants were re-potted in 30 cm pots using the same soil mixture with the addition of Osmocote Exact Standard 3-4M (ICL) slow release fertilizer. Plants were arranged randomly for the duration of growth to eliminate greenhouse location effects.

| Chlorophyll extraction and measurements
A 0.25 cm 2 leaf disk was taken from ¾ along the proximal-distal axis of leaf 4 23 days after planting, avoiding the midrib. The leaf disk was frozen in liquid nitrogen, homogenized and resuspended in 2 ml of ice-cold 80% (v/v) acetone before centrifugation (17900 g) for 5 min at 4°C. The supernatant was then removed and kept on ice in the dark prior to measurements. 1 ml of supernatant was transferred to a 1.5 ml quartz cuvette and absorbance recorded at 646.6, 663.6, and 750 nm. Total chlorophyll, chlorophyll a and chlorophyll b contents were calculated on a leaf area basis as described previously (Porra, Thompson, & Kriedemann, 1989).

| Quantitative RT-PCR analysis
Two 0.25 cm 2 leaf disks per plant were harvested from developmentally equivalent immature leaves emerging from the whorl 31 days after planting, and frozen in liquid nitrogen. Total RNA was extracted using an RNeasy Plant Mini Kit following the manufacturer's instructions (Qiagen). RNA was then treated with DNaseI (Invitrogen), and 1 µg of DNase-treated RNA was used as template for cDNA synthesis using a Maxima First Strand cDNA synthesis kit (Thermo Scientific). cDNA quality was assessed by RT-PCR using primers amplifying a product with distinct sizes in cDNA and genomic DNA extracts.
Quantitative RT-PCR (qRT-PCR) experiments were undertaken using SYBR Green. Cycle conditions were 95°C for 10 min, then 40 cycles of 95°C for 15 s and 60°C for 1 min. Melt curves were generated between 60°C and 95°C to confirm a single product was amplified ( Figure S1). Primers were designed for the CDS of both ZmGLK1 (GRMZM2G026833) and ZmG2 (GRMZM2G087804) using Primer3Plus (Untergasser et al., 2012), amplifying regions ~100 bp in length ( Figure S1a). Primers were first tested by RT-PCR using Gotaq (Promega) with cycle conditions 95°C for 5 min, 35 cycles of 95°C for 30 s, 57°C for 30 s and 72°C for 30 s, and 72°C for 5 min, to confirm correct amplicon size ( Figure S1b).
qRT-PCR amplification efficiencies were calculated using 1/4, 1/8, 1/16, 1/32 and 1/64 wild-type cDNA dilutions. Experiments were run with primers for two housekeeping genes, ZmCYP and ZmEF1α, that have been previously validated as suitable controls in maize leaf experiments (Lin et al., 2014). Three technical replicates were undertaken for each sample and confirmed to have Ct values with range <0.5. All relevant comparisons were run on the same plate alongside water controls, and as such the WT (m2m1) samples were repeated alongside the m2m2 samples. Ct values were calculated using the real-time PCR miner algorithm (Zhao & Fernald, 2005), and fold-change values were calculated using the 2 −ΔΔCT method (Livak & Schmittgen, 2001). Each wild-type sample was compared individually to the average wild-type values in order to indicate the spread of the wild-type data. Mutant samples were then compared to the same wild-type average.

| Single cell isolation and chloroplast quantification
Strips of leaf ~1 cm long and ~1-2 mm wide were cut from the midpoint along the proximal-distal axis of fully expanded leaf 5, 33 days after planting. Leaf strips were harvested and fixed in 0.5% glutaraldehyde in PBS (0.15 M NaCl, 10 mM phosphate buffer pH 6.9), for 1 hr in the dark at room temperature as previously described (Wang et al., 2017). Samples were then stored in Na 2 EDTA buffer (0.2 M disodium EDTA, pH 9) at 4°C until digestion. For single cell isolation, samples were incubated in Na 2 EDTA buffer for 3 hr at 55°C, before being rinsed in digestion buffer (0.15 M sodium hydrogen phosphate, 0.04 M citric acid, pH 5.3) twice for 15 min. Samples were then incubated at 45°C in 2.5% pectinase in digestion buffer for 1 hr, before rinsing twice in digestion buffer. All digested samples were stored at 4°C and imaged within a week.
For imaging, digested leaf strips were placed on a glass slide in a drop of 1:1 Na 2 EDTA buffer: glycerol. The bottom of a 1.5 ml eppendorf tube was used to apply pressure to the leaf strip in order to separate the cells. For the isolation of single M cells, light pressure was applied by tapping the tube bottom on the edges of the leaf strip, whereas more force was applied for the isolation of single BS cells. Images were acquired using a Leica SP5 laser scanning confocal microscope and Leica Application Suite (LAS) software. Chlorophyll was excited at 561 nm and emission collected between 650 nm and 750 nm. Transmitted laser light provided pseudo bright field images. Cells were imaged in each focal plane at 500 nm step size such that a z-stack comprising the entire cell was generated.
Image quantification was undertaken using Fiji (https://imagej. net/Fiji). M cell volumes were calculated by assuming a cuboid shape, whereas BS cells were treated as cylinders. M chloroplast volumes were calculated using the volumest plugin in Fiji. M cell chloroplast counts were undertaken blind such that the genotypes were unknown until after quantification was complete. Total BS chloroplast volumes were calculated by thresholding the image using the Otsu method. Minimum and maximum threshold parameters were automatically calculated, although in some cases manual adjustment was necessary to ensure all chloroplasts were included. The area of threshold pixels was then counted in each z-stack layer before being combined for an estimated total chloroplast volume per cell.

| Transmission electron microscopy
Leaf 4 was sampled with a 2 mm leaf punch at the mid-point along the proximal-distal axis 29 days after planting. Leaf disks were cut in half and then fixed in a microwave fixation unit (Leica EM AMW) with 2.5% glutaraldehyde 4% paraformaldehyde in 0.1 M sodium cacodylate buffer, pH 6.9, for 4 × 2 min at 37°C, alternating 20 W and 0 W microwave power (continuous), before final incubations for 2 min with 30 W (pulse), 2 min with 0 W power and 4 min with 20 W (pulse), all at 37°C. Samples were then placed in fresh fixative at 4°C for 3 days. Unless stated all subsequent microwave power settings were continuous rather than pulsed.
Sections were post-stained with 2% lead citrate for 5 min before rinsing with degassed dH 2 O four times. Sections were then dried and imaged using a FEI Tecnai 12 transmission electron microscope operated at 120 kV with a Gatan OneView camera.

| Photosynthesis measurements
Steady state A max measurements and A/Ci curves were generated using an infrared gas analyser (Li-6800, LI-COR) with a 2 cm 2 leaf chamber fluorometer head. A max measurements were undertaken in the greenhouse on fully expanded leaf 8, 49 days after germination, at 800 μmol/mol CO2, 1800 μmol photons m −2 s −1 , 60% humidity and 26°C leaf temperature. Leaves were clamped at the midpoint along the proximal-distal axis, adjacent to but not including the midrib. Leaves were allowed to acclimate for around 30 min until photosynthetic rate had stabilized, at which point A max was recorded. Measurements were taken throughout the day between 10:00 and 18:00, with genotypes randomly sampled to avoid any circadian influences. A/C i curves were generated using fully expanded leaf 6 between 29 and 32 days after planting. Plants were moved into the lab where leaves were clamped in the same way as for A max measurements and allowed to acclimate for at least 30 min at 400 μmol/mol CO2, 1800 μmol photons m −2 s −1 , 60% humidity and leaf temperature 26°C. A/C i curves were generated by measurement of photosynthetic rate at 400, 300, 200, 100, 50, 25, 400 (recovery, omitted from curve plotting), 600, 800, and 1000 μmol/ mol CO 2 . At each new CO 2 concentration, measurements were only taken at steady state, normally achieved between 2 and 6 min after the desired CO 2 concentration was achieved.
Individual A/C i curves were modelled using the equations described previously for C 4 photosynthesis (von Caemmerer, 2000).
The most recent kinetic constants described for Setaria viridis (also an NADP-ME C 4 monocot) were used in order to fit the model and estimate both V cmax and V pmax (Boyd, Gandin, & Cousins, 2015).
Modelled data agreed closely with experimental data. Compensation points were estimated by fitting a simple polynomial trendline to the first 3-4 points of the modelled data, and solving this to estimate the x-axis intercept and thus the CO 2 compensation point.

| δ 13 C‰ analysis
A small section of leaf tissue was sampled from leaf 10, 58 days after planting. This tissue was then freeze dried for 1 week. Once dry, tissue was ground to a powder by milling in a TissueLyser II (Qiagen) for 1 min at 28 Hz. From this powder, 1 mg of tissue was weighed and placed in a tightly sealed tin capsule. 13 C/ 12 C ratios were measured using a Sercon 20/22 isotope-ratio mass spectrometry system coupled to a Sercon GSL elemental analyser. Samples were run alongside collagen standards which are traceable back to the VPDB standard.
Known alanine controls were run alongside samples to ensure accuracy. δ 13 C ‰ was calculated as: where R is the ratio of 13 C mass to 12 C mass.

| Chlorophyll levels are reduced in Zmscr1;Zmscr1h mutants
Zmscr1;Zmscr1h plants appear paler green than either wild-type or single mutants. To quantify this difference, the chlorophyll content in multiple individuals from each genotype was measured (Figure 1). In general, overall chlorophyll levels in WT (m2m1) (a segregating wildtype line from the m2m1 F2 population) and single mutant leaves were similar, although levels were slightly lower in the Zmscr1h-m2 line (Figure 1a). In contrast, both Zmscr1-m2;Zmscr1h-m1 and Zmscr1-m2;Zmscr1h-m2 mutants (hereafter referred to as the m2m1 and m2m2 lines for simplicity) had significantly lower levels of total chlorophyll than WT on a leaf area basis (Figure 1a). To ensure that this difference was not indirectly caused by a reduction in leaf thickness, the thickness of leaves was quantified in both double mutants (Hughes et al., 2019).
A slight mean reduction in the thickness of the m2m1 leaves was observed, although it was smaller in magnitude and less consistent than the chlorophyll reduction, and there was no change in the m2m2 line ( Figure S2). Altered leaf thickness does not therefore account for the reduction in chlorophyll. When analysed individually, both chlorophyll a and chlorophyll b levels were reduced by a similar magnitude in Zmscr1;Zmscr1h mutants (Figure 1b,c). The pale green phenotype of Zmscr1;Zmscr1h double mutants is thus caused by a 52% (m2m1) or 42% (m2m2) reduction in total chlorophyll levels, a defect that is complemented by activity of either ZmSCR1 or ZmSCR1h in single mutants. All subsequent analyses therefore focussed on comparisons between wild-type and double mutant plants. ZmG2 transcripts accumulate in BS cells (Rossini, Cribb, Martin, & Langdale, 2001). Loss of ZmG2 function specifically perturbs BS chloroplast development (Hall, Rossini, Cribb, & Langdale, 1998;Langdale & Kidner, 1994). We have previously shown that ZmSCR1

| ZmGLK1 and
and ZmSCR1h transcripts accumulate most strongly in the ground meristem cells of leaf primordia that give rise to M cells in the mature leaf (Hughes et al., 2019). We therefore hypothesized that if ZmSCR1 and/or ZmSCR1h had a direct effect on chloroplast development, ZmGLK1 but not ZmG2 transcript levels would be reduced in Zmscr1;Zmscr1h mutant leaf primordia. Consistent with this hypothesis, qRT-PCR analysis revealed that ZmGLK1 transcript levels were significantly reduced by on average 62% (m2m1) or 72% (m2m2) in double mutants (Figure 2a). Although there was some biological variation between individuals of both wild-type and mutant lines, this pattern was consistent and was replicated in two independent experiments. By contrast, levels of ZmG2 transcripts were only significantly different from wild-type in the m2m1 line, and even in this line the reduction was far less severe (reduced by only 43%) than seen with ZmGLK1 (Figure 2a). Together these results suggested a direct or indirect role for ZmSCR1 and ZmSCR1h in the regulation of ZmGLK1, and thus possibly in the development of M cell chloroplasts.

| Chloroplast development is largely unaltered in Zmscr1;Zmscr1h mutants
To assess whether reduced ZmGLK1 transcript levels in Zmscr1;Zmscr1h double mutants led to perturbed chloroplast

| Zmscr1;Zmscr1h operate a normal C 4 photosynthetic cycle with reduced capacity
Given that growth and chlorophyll levels are perturbed in Zmscr1;Zmscr1h mutants but chloroplast development is largely unaffected, we sought to determine whether photosynthetic capacity was compromised. As an initial screen, the maximal rate of carbon assimilation (A max ) was measured under saturating CO 2 conditions. Figure 3a shows that in both Zmscr1;Zmscr1h mutant lines, A max was on average around 20 µmol/m 2 s −1 as compared to around 30 µmol/ m 2 s −1 in wild-type. This reduction was consistent across all plants measured (n = 6 for each genotype), across two independent experimental set-ups ( Figure S4). To test whether Zmscr1;Zmscr1h mutants also had reduced photosynthetic capacity under atmospheric conditions (400 μmol/mol CO 2 ), A/C i (CO 2 assimilation, A, versus substomatal CO 2 concentration, C i ) curves were generated for at least three different plants from each genotype (Figure 3b). At 400 μmol/ mol, C i levels were similar between genotypes, suggesting that stomatal behaviour was not significantly altered ( Figure 3b). However, stomatal conductance at each CO 2 concentration was on average lower in the double mutants ( Figure S5). Thus, the apparently normal C i values may be underpinned by a reduction in CO 2 utilisation in the mutants. In support of this, at 400 μmol/mol CO 2 , photosynthetic rates were reduced by a similar magnitude to that seen at saturating CO 2 , with rates only slightly below the A max rate in both wild-type and mutants (Figure 3b). This reduction in photosynthetic capacity likely underpins the perturbed growth phenotype in Zmscr1;Zmscr1h mutants.
In mature leaves of both Zmscr1;Zmscr1h mutant lines, the majority (>50% in both cases) of veins are separated by only one M cell rather than the normal two, skewing the 1:1 ratio of BS:M cells assumed to be critical for efficient C 4 photosynthesis in wild-type maize and other Kranz exhibiting C 4 species (Hughes et al., 2019). To determine whether the relative reduction in M cell volume leads to reduced PEPC-mediated carbon fixation capacity in Zmscr1;Zmscr1h mutants, we modelled the A/C i curves using previously established C 4 photosynthesis equations (von Caemmerer, 2000). These modelled assimilation curves enabled the estimation of two key parameters, the maximal rate of RuBisCO carboxylation (V cmax ) and the maximal rate of PEPC carboxylation (V pmax ). Figure 3c shows that V cmax was reduced in both mutant lines, consistent with the lower maximal rates of photosynthesis observed. V pmax estimation, which is based on the initial slope of the A/C i curve, was also reduced in the mutants compared to the wild-type (Figure 3d). The estimated V pmax in Zmscr1;Zmscr1h mutants was around a third of that found in wild-type (Figure 3d), indicating a significant reduction in PEPC capacity at a whole leaf level.
Given the reduction in PEPC capacity in Zmscr1;Zmscr1h mutants, alongside the perturbed Kranz patterning, it seemed plausible that some CO 2 could be fixed directly by RuBisCO in the BS in a C 3 cycle. To test this, we extrapolated A/C i curves to estimate the CO 2 compensation point in wild-type and Zmscr1;Zmscr1h plants.
This analysis revealed no significant difference in the CO 2 compensation point between wild-type and Zmscr1;Zmscr1h mutants ( Figure 3e). We also carried out a δ 13 C ‰ analysis, which exploits the fact that PEPC and RuBisCO discriminate differently between 12 C and 13 C. Fixation by PEPC in C 4 species results in δ 13 C ‰ values between −12 and −16, whereas fixation by RuBisCO in C 3 species gives values between −20 and −37 (O'Leary, 1981).
Notably, consistent with the CO 2 compensation point estimations,

| D ISCUSS I ON
The coordination of growth and development are of fundamental importance throughout the plant lifecycle. SCARECROW is one of the best studied plant developmental regulators, with roles in patterning processes in both roots and shoots (Cui et al., 2014;Di Laurenzio et al., 1996;Wysocka-Diller et al., 2000). However, comparatively little attention has been paid to the perturbations in growth and physiology which accompany patterning defects in loss of function  (Ma et al., 2014), and protochlorophyllide oxidoreductase B, which catalyses one of the latter steps in chlorophyll biosynthesis, is a putative direct target of AtSCR itself (Cui et al., 2014;Garrone, Archipowa, Zipfel, Hermann, & Dietzek, 2015). AtSCR also directly modulates the sugar response (Cui, Hao, & Kong, 2012) and genes encoding components of the light harvesting complex of photosystems I and II are putative targets (Cui et al., 2014). Most interestingly, AtSCR is proposed to bind to the promoter of AtGLK2 but not AtGLK1 (Cui et al., 2014). In light of these data, and given the functional conservation between maize and Arabidopsis SCARECROW (Lim et al., 2005), it is plausible that ZmSCR1/1h directly regulates expression of ZmGLK1 and/or ZmG2.
In Zmscr1;Zmscr1h mutants, ZmGLK1 transcript levels are reduced to a greater extent than ZmG2 levels on a whole leaf basis. Black lines connect the lowest and highest value in each genotype. Statistical significance between WT (m2m1) and mutants was assessed using a one-way ANOVA: *p < .05; **p < .01; ***p < .001; n.s. p ≥ .05

F I G U R E 3 Photosynthesis is perturbed in
This observation is consistent with the suggestion that ZmSCR1 could directly activate ZmGLK1 gene expression because ZmSCR1 and ZmSCR1h are expressed predominantly in M cell precursors (Hughes et al., 2019). Some of the reduction in ZmGLK1 levels can be attributed to the altered M cell patterning that results in >50% vascular bundles being separated by only one M cell rather than the normal two. However, this anatomical difference could only explain a 25% reduction in transcript levels at the whole leaf level whereas the observed reduction was 62% (m2m1) or 72% (m2m2). As such there is a ~35%-40% reduction in ZmGLK1 transcript levels that is independent of any patterning defects, a similar reduction to that seen for ZmG2 transcripts. Given that ZmG2 is expressed preferentially in BS cells, the most parsimonious explanation for this observation is that loss of ZmSCR1/ZmSCR1h function has an indirect effect on ZmGLK1/ZmG2 expression, although the possibility of direct but non-cell autonomous activation cannot be excluded. In either case, it can be concluded that ZmGLK1 and ZmG2 levels are high enough in Zmscr1;Zmscr1h mutants to facilitate normal chloroplast development.
Given the marked reduction in total chlorophyll in Zmscr1;Zmscr1h mutants (~50% of wild-type), it is counter-intuitive that chloroplast development is apparently normal because intermediates in the chlorophyll biosynthetic pathway are known to participate in retrograde signalling pathways from the chloroplast to the nucleus (Larkin, 2016;Larkin, Alonso, Ecker, & Chory, 2003;Mochizuki, Brusslan, Larkin, Nagatani, & Chory, 2001). It is possible that ZmSCR1 and ZmSCR1h affect chlorophyll biosynthesis at a point in the pathway downstream of the intermediates responsible for retrograde signalling and the suppression of chloroplast development.
However, this is unlikely given the reduction in ZmGLK1 and ZmG2 transcript levels, because the Arabidopsis orthologs AtGLK1 and AtGLK2 directly activate chlorophyll biosynthesis genes throughout the pathway (Waters et al., 2009). An alternative possibility is that ZmSCR1 and ZmSCR1h play a direct or indirect role in the retrograde signalling pathway itself, such that loss of function prevents feedback from the chloroplast to the nucleus. In either case, our results indicate that at the level of chlorophyll reduction observed, it is possible to uncouple chlorophyll biosynthesis and chloroplast development.
Although both chloroplast development and the C 4 photosynthetic signature appear normal in Zmscr1;Zmscr1h mutants, lower chlorophyll levels are accompanied by reduced maximum photosynthetic capacity (A max ), RuBisCO carboxylation rate (V cmax ) and PEPCcarboxylation rate (V pmax ). The reduction in Zmscr1;Zmscr1h stomatal conductance is interesting in light of the recent finding that SCR orthologs in rice regulate stomatal patterning . It is possible that despite distinct transcript accumulation patterns (Hughes et al., 2019;Kamiya, Itoh, Morikami, Nagato, & Matsuoka, 2003), SCR also plays a role in stomatal patterning in maize. However, even if altered gas exchange contributes to the compromised photosynthetic performance observed in Zmscr1;Zmscr1h mutants, it is unlikely to be a major factor because photosynthetic capacity is not only reduced in atmospheric CO 2 but also at saturating CO 2 concentrations, where stomatal conductance no longer limits photosynthesis. Reduced V pmax may be a reflection of the relative reduction in total M cell volume (and thus reduction in relative M volume in the gas-exchange chamber) that results from the known patterning defect. However, the reduction in V cmax is unlikely to result from anatomical changes because BS cells, and the chloroplasts within them, are unaltered; there is in fact a relative increase in the proportion of BS cells due to the M patterning defect (Hughes et al., 2019). It is more likely that the observed reduction in overall photosynthetic capacity results from a reduction in the amount of light that can be harvested to power photosynthesis as a consequence of the lower chlorophyll levels. Collectively, our results demonstrate that in addition to well-established patterning roles, SCR helps to establish and/ or maintain photosynthetic capacity in maize.

DATA AVA I L A B I L I T Y S TAT E M E N T
All relevant data is included in the manuscript and supporting information. Access to raw data is available on request.