Greater Than the Sum of the Parts: Revisiting the Enhancement Effect in Photosynthesis Using Simulated Sun‐ and Shade‐Light

ABSTRACT

Far‐red light (FR) alone drives photosynthesis poorly, but when combined with shorter wavelengths it enhances photosynthesis beyond the sum of their individual effects—a phenomenon known as the Emerson enhancement effect. This effect is well‐established for narrowband PAR‐FR mixtures, and recent results show it also occurs within broadband “white” light, though the spectra investigated differed from those observed under natural conditions. In this study, we used simulated sun and foliar shade spectra (SUN and SHADE) during growth and measurements of mature tomato leaves to determine quantum yield for CO₂ assimilation (Φ_CO2) on an absorbed light basis under SUN and SHADE, 17 narrowband irradiances, and combinations of SUN or SHADE with the 17 narrowband irradiances. Enhancement was calculated for each of the 53 unique spectra by comparing predicted and measured Φ_CO2. This study shows that a 23% enhancement occurs in the simulated SHADE spectrum light and involves the whole spectrum, and in the simulated SUN spectrum enhancement is absent. Further enhancement was observed when narrowband irradiances were added to the SUN or SHADE spectra. The exclusion of the far‐red region (> 700 nm) by PAR‐based light intensity measurement is particularly problematic in natural, far‐red‐rich, canopy environments where far‐red has surprising photosynthetic utility.

Summary statement

• •In this paper we revisited the photosynthetic enhancement effect on the light‐use efficiency of carbon dioxide fixation of tomato (Lycopersicum esculentum) plants grown under simulated sun and shade light, using simulated sun and shade light and 17 narrowband irradiances as the measuring light.

Abbreviations

ATP

adenosine triphosphate

Fd

ferredoxin

Fm

dark‐adapted maximum fluorescence

Fm'

light adapted maximum fluorescence

Fo

dark‐adapted minimum fluorescence

Fo'

light‐adapted minimum fluorescence

Fv'/Fm'

maximum light adapted efficiency of PSII

LHCII

light‐harvesting complex II

NADPH

nicotinamide adenine dinucleotide phosphate

P700

photosystem I primary donor

PAR

photosynthetically active radiation

PQ

plastoquinone

PSI

photosystem I

PSII

photosystem II

Q_A

primary quinone acceptor of PSII

SE

standard error

Φ_CO2

quantum yield of CO₂ fixation

Φ_PSII

quantum yield of electron transport by PSII

1. Introduction

The photosynthetically active wavelengths used by plants are commonly defined to extend over a narrow wavelength region (400–700 nm) of the total shortwave solar radiation that reaches the Earth's surface, which itself is generally defined as ranging from 280 to 3000 nm. This definition of ‘photosynthetically active’ is a practical definition rather than one based on the biophysical limits of photosynthesis. Extensive study of the spectral dependence of photosynthesis in leaves on an absorbed light basis has shown red light drives photosynthesis most efficiently followed by short‐wave blue light (400–420 nm), and that there is a minimum in the long‐wave blue region at around 480–500 nm (Hogewoning et al. 2012; Inada 1976; McCree 1972a). Invariably, photosynthetic efficiency falls sharply above ~680 nm and rapidly diminishes to zero as wavelength increases in a phenomenon first identified in Chlorella and termed the ‘red drop’ (Emerson and Lewis 1943). These observations can easily lead to the conclusion that far‐red light can be disregarded in photosynthesis measurements and in practice PAR is often restricted to irradiance between 400 and 700 nm. However, light in natural environments comprises PAR and far‐red, and a conspicuous feature of far‐red light is that its addition to PAR commonly leads to a rate of photosynthesis which generally exceeds what could be predicted by simply adding photosynthetic rates for each of the far‐red and PAR spectra together; that is, the rate of CO₂ fixation is greater than the sum of the parts. This synergistic effect, eponymously termed the Emerson enhancement effect (hereafter referred to as ‘enhancement’), was discovered by presenting cells of Chlorella pyrenoidosa with PSI (700 nm) and PSII light (680 nm), applied separately and simultaneously (Emerson 1957; Emerson 1958). This led in turn to the discovery that there are two systems operating in series (Emerson et al. 1957; Duysens et al. 1961), known today as ‘photosystem 1'and ‘photosystem 2’; whereas the PAR wavelengths mostly over‐excite PSII compared to PSI, far‐red light excites PSI considerably more strongly than PSII (Hogewoning et al. 2012; Mattila et al. 2020). The spectral qualities of light which preferentially excite PSI or PSII are often called simply PSI or PSII light (Chow 1990).

A prerequisite for enhancement is the balanced–or better balanced–turnover of the two photosystems. The predominant form of electron transport in C3 plants is linear electron transport (LET) which depends on the cooperative action of photosystems I (PSI) and II (PSII) working in series with equal electron transport rates through PSI and PSII. Since the two photosystems have different light‐absorption spectra, the spectral composition of irradiance can result in the over‐excitation of one photosystem or the other. Furthermore, maximum light‐use efficiencies for electron transport differ between PSI and PSII and this must be taken into account when considering the balance between the photosystems: whereas PSI operates with an efficiency of ~0.99, PSII is less efficient with a maximal operating efficiency of about ~0.9 (Pfündel 1998). If the potential electron transport rate through one photosystem is greater than that through the other, the system with excess capacity will be limited by the other and lose efficiency until the capacities are in balance. The operation of cyclic electron transport (CET) around PSI could also result in limitation of PSII electron transport as PSI competes with PSII for the electron acceptors that would normally be reduced by PSII in LET. Under light‐limiting conditions the limitation of one or the other photosystem by the other in this way can result in the loss of overall photosynthetic light‐use efficiency at the level of leaf assimilation (Hogewoning et al. 2012). Clearly, balanced photosystem turnover and balanced activity of CET and LET is the product of numerous factors which cannot be viewed in isolation.

The importance of LET and high photosynthetic efficiency is evidenced by the ability of leaves to acclimate to the spectrum in which they were produced to maximise LET. For example, leaves produced under PSI light have a greater PSII:PSI ratio than leaves produced under PSII light, whereas this ratio is lower in leaves produced in PSII light (Chow et al. 1990; Walters and Horton 1994; Hogewoning et al. 2012). Pea leaves grown in PSI or PSII irradiance and subsequently measured with each of those irradiances showed that Φ_CO2 on an absorbed light basis was highest when spectra were the same, demonstrating the ‘memory’ of leaves caused by this acclimation (Chow et al. 1990). Even after full leaf expansion has been achieved, the plasticity of the photosynthetic apparatus arising from acclimatory mechanisms further highlights the importance to the plant of maintaining a high photosynthetic efficiency. These acclimatory mechanisms have been divided into short‐term and long‐term mechanisms based on the duration which they require to take effect. The short‐term state transitions occur within minutes in response to over‐excitation of one of the photosystems and involve the relocation of a mobile fraction of LHCII from PSII to PSI, or vice versa (Croce 2020), to increase cross‐section and absorption of the rate‐limiting photosystem (Taylor et al. 2019). State transitions have been shown to have a direct beneficial effect on photosynthetic efficiency through this structural re‐organisation, which at least partially mitigate excitation imbalance (e.g. Taylor et al. 2019). A longer‐term acclimation mechanism, occurring in the order of days, is the adjustment to the stoichiometry of PSI and PSII which involves the synthesis and degradation of various PSII and PSI proteins (Kim et al. 1993).

The majority of early work on enhancement used narrowband irradiances to explore and characterise enhancement in algal cells (e.g. (Govindjee and Rabinowitch 1960; Myers and Graham 1963; Senger and Bishop 1969), isolated chloroplasts (Govindjee et al. 1964), or intact leaves (Canaani et al. 1982). Though the use of narrowband irradiances in these seminal studies simplified the characterisation of enhancement, comparatively little attention was given at the time to enhancement within broad‐spectrum or white light. In spinach chloroplasts the addition of a range of narrowband red and far‐red wavelengths (678–740 nm) to background white light revealed significant enhancement in terms of NADP reduction in the far‐red region, increasing up to the longest narrowband wavelength used (Govindjee et al. 1964). The significant ‘enhancement’ observed by those authors when using supplementary wavelengths as long as 740 nm indicates the photosynthetic relevance of far‐red light well outside the PAR region for which Φ_CO2 is zero (or nearly so) when those wavelengths are used alone (e.g., in cucumber leaves illuminated with 736 nm; Hogewoning et al. 2012).

An investigation using leaves of selected crop species and four types of ‘white’ light, including a spectrum rich in far‐red light from a quartz‐iodine lamp, concluded that enhancement could be ignored in white light (McCree 1972b). However, enhancement in white light has recently been revisited with emphasis on the role of far‐red light in this phenomenon (Zhen and van Iersel 2017; Zhen and Bugbee 2020; see also Huber et al. 2024). In lettuce leaves subjected to 400 µmol m^{− 2} s^{− 1} cool‐white LED light, the addition of 60 µmol m^{− 2} s^{− 1} far‐red light (with a peak wavelength of 735 nm) resulted in comparable canopy photosynthetic rates to when 60 µmol m^{− 2} s^{− 1} more of the same white light was added (total PAR of 460 µmol m^{− 2} s^{− 1}), despite the far‐red light being only weakly absorbed compared to the white light (Zhen and Bugbee 2020). This led those authors to conclude that the added far‐red irradiance was utilised with greater efficiency than the white light on an absorbed light basis, indicating that significant enhancement had occurred. More recently, the addition of 500 µmol m^{− 2} s^{− 1} far‐red light to a white light background (comprising only a low proportion of transmitted irradiance in the greenhouse supplemented with white LED light to a PAR of 400 µmol m^{− 2} s^{− 1}) led to significant enhancement in selected greenhouse‐grown rice varieties (Huber et al. 2024).

The quantitative assessment of the role of far‐red light (or any other spectral region) in enhancement is easier when using strictly light‐limiting intensities on an absorbed light basis. Such an approach avoids spectral additions extending into nonlinear portions of the light response curve of photosynthesis, as well as differences in absorption between the added light and background light; together these create difficulties in quantifying potential enhancement. Also of importance is the spectral composition of the white‐light spectrum used. Apart from the work of Hogewoning et al. (2012) in cucumber, the use of more natural balanced sun‐ or shade‐like PAR and far‐red spectra as growth and actinic background ‘white’ irradiance in far‐red enhancement‐related work is notably absent in literature. This limits the understanding of the potential significance of enhancement in natural ecophysiological contexts where the spectrum can change alongside fluctuations in intensity, as well as in practical settings such as in high‐tech greenhouses when supplementary LED lighting is added to natural daylight. The focus of this study was to test for and quantify enhancement in simulated sun‐ and shade‐light using leaves produced under each of these spectra to create more natural, yet contrasting, spectral acclimation histories. Chlorophyll fluorescence and 820 nm absorbance change measurements are also undertaken to quantitatively assess accompanying thylakoid‐level changes.

2. Materials and Methods

2.1. Plant Materials and Growth Conditions

Tomato (Solanum lycopersicum cv. Moneymaker) seeds were sown in rockwool blocks (Grodan, Roermond, The Netherlands) and grown for 30–35 days under a simulated sunlight spectrum (‘SUN’), or a simulated shade light spectrum (‘SHADE’) in two light‐separated cabinets (80 cm × 80 cm × 50 cm). The cabinets were made of opaque polythene sheets and positioned within one large climate‐controlled room which was kept at 20°C air temperature and 70% relative humidity. The rockwool blocks were flooded with Hoagland's solution for 15 min daily. During the 16‐h photoperiods, PAR was maintained at 100 µmol m^{− 2} s^{− 1} at the apical bud until the third leaf began to develop, at which point the third leaf became the reference point for PAR maintenance. The temperature of the third leaf during growth was controlled at 22.5°C by adjusting the flow rate of ambient air from the large climate room into the smaller growth cabinets using a closed control loop including continuous leaf temperature monitoring by a thermocouple appressed to the leaf. After 30–35 days the third, fully expanded leaf was used for all photosynthesis measurements.

2.2. Growth‐ and Actinic Light

The SUN and SHADE growth spectra were identical to those shown in Taylor et al. (2019). A sulphur plasma lamp (Plasma International, Muhlheim am Main, Germany) provided the simulated SUN spectrum light during growth, whereas quartz‐halogen lamps, filtered with a plastic‐film filter (Full C.T. blue, Lee Filters, Hampshire, U.K.), provided the simulated SHADE spectrum light during growth. The SUN growth spectrum comprised more short wavelength PAR than the SHADE growth spectrum, which was rich in far‐red. Actinic SUN light for photosynthesis measurements was produced by a xenon lamp (Newport Instruments, USA) filtered by a Calflex NIR filter (Balzers, Lichtenstein). The spectral output of this lamp closely matched a standard ASTM solar spectrum (Figure 1B; G173, ASTM, 2003). The same xenon lamp combined with plastic sheet filters (‘Yellow’ and ‘Cold Blue’; Lee Filters, Hampshire, UK) and a Calflex NIR filter was used to produce actinic SHADE light. In terms of the far‐red composition of the actinic SUN and SHADE spectra, the percentage of total photon flux in the band 700–728 nm was 10% and 28%, respectively. Although these simulated actinic sun‐ and shade‐light spectra were matched as closely as was possible to their respective natural reference spectra, they are deficient in the shorter wavelengths of the blue region and were cut off above about 740 nm. In the case of the shade spectrum, a further difference is a short‐wave shifted far‐red maximum (relative to a measurement of spectrum under a woodland canopy; Figure 1C). For these reasons, neither spectrum can be strictly defined as either sunlight or shade light. However, the spectra are distinctly sun‐like or shade‐like, and for the purposes of this study these spectra are referred to as ‘SUN’ and ‘SHADE’.

Figure 1.

Relative quantum fluxes of the actinic spectra used for photosynthesis measurements: narrowband spectra (A), simulated sun spectrum (B, solid line), and simulated shade spectrum (C, solid line). Dotted lines in (B) and (C) represent, respectively, a standard solar spectrum (G173, ASTM, 2003) and a shade spectrum measured beneath a woodland canopy during summer in the Netherlands. Actinic spectra also included the combination of each narrowband spectrum with either the simulated SUN or SHADE spectrum in a 1:1 ratio on an absorbed light basis.

In addition to the actinic SUN and SHADE light spectra, photosynthesis was also measured in 17 narrowband actinic light spectra, alone, and in combination with, actinic SUN and SHADE light. Narrowband actinic light was produced by LED light sources, which had nominal peak wavelengths of 406, 427 and 445 nm, and from 460 to 720 nm peak wavelengths that progressed in 20 nm increments (Figure 1A). These irradiances were produced using a laboratory‐built light source comprising LEDs mounted on an optical bench. Light from a high‐power white LED was filtered using bandpass interference filters to produce the following narrowband spectra: 460 nm with 10 nm full width at half maximum (FWHM; Thorlabs, Newton, NJ, USA) and 500 to 680 nm with 10 nm FWHM (Thorlabs with the exception of 500 and 520 nm where Edmund filters were used; Edmund Optics, Barrington, NJ, USA). In addition to the bandpass filters used in combination with the white LED, a Calflex filter and one of four dichroic filters (DT blue, DT green, DT Yellow or DT Red as appropriate, Balzers, Liechtenstein) were used for thermal management and blocking of longwave transmission by some filters. The remaining narrowband spectra were produced using single‐colour LEDs in combination with bandpass filters: UV LEDs for 406 nm with 15 nm FWHM 406 nm interference filter (Semrock, Rochester, NY, USA), 420 nm LEDs (SemiLed, Taiwan, ROC) for 427 nm with 15 nm FWHM 427 nm interference filter (Semrock), 440 nm LEDs for 445 nm with 15 nm FWHM 445 nm interference filter (Semrock), 480 nm LEDs (Luxeon Rebel, Luxeon, San Jose, CA 95131, USA) for 480 nm with a 10 nm FWHM interference filter, 700 nm LEDs (Roithner Lasertechnik, Vienna, Austria) for 700 nm with a 10 nm FWHM 700 nm interference filter (Thorlabs), and 720 nm LEDs (Roithner) for 720 nm with a 720 nm 10 nm FWHM interference filter (Thorlabs). No CO₂ fixation was detected when using 736 nm light and therefore longer wavelengths were not used. This 736 nm light was produced using a far‐red LED and a 740 nm interference filter (Thorlabs).

2.3. Gas Exchange System

The measurement of spectral enhancement of carbon dioxide fixation and its correlation with the light‐use efficiency and regulation of photosystems I and II required the concurrent measurement of assimilation, chlorophyll fluorescence, and near‐infra‐red light‐induced absorbance changes, in a leaf chamber that also allows full control and accurate measurement of the spectral irradiances. To achieve this, we used a custom two‐part leaf chamber (Supporting Information S1: Figure S1) that has been partly described in Hogewoning et al. (2012) and updated by Taylor et al. (2019). The gas mix used for measurements comprised 400 µmol mol^{− 1} CO₂, 210 mmol mol^{− 1} O₂, and 18.8 mmol mol^{− 1} H₂O with the remaining fraction being N₂. The gas stream was split into reference and sample streams in which a LI‐7000 CO₂/H₂O analyser (LI‐7000, Li‐Cor, Nebraska, USA) analysed each stream. Sample and reference streams were routinely cross‐checked before measurement using nitrogen (for 0 µmol mol^{− 1} CO₂) and a reference gas comprising 391 µmol mol^{− 1} CO₂, and regularly cross‐checked against another calibrated gas analyser (Li‐6400, Li‐Cor, Nebraska, USA). CO₂ readings agreed within 0.1% across a range of CO₂ concentrations. The flow rate was maintained at 250 cm³min^{− 1} using a mass‐flow controller and cross‐checked against another flow controller used as a flow metre, which agreed with 0.2% margin. This second flow metre was also used to test for leaks by monitoring the exhaust flow from the leaf chamber. Leaks were reliably eliminated by using a quick‐set nontoxic silicone rubber (Body Double, Smooth‐On, Macungie, PA, USA) between the abaxial leaf surface and the lower chamber seal. Leaf temperature was maintained at 22.5°C and monitored using a calibrated noncontact temperature sensor (Micro Irt/c, Exergen, Watertown, MA, USA) mounted directly below the leaf. Leaf temperature was controlled by circulating water from a temperature‐controlled water bath through an internal channel in the upper and lower leaf‐chamber halves.

Actinic light was delivered to the leaf chamber by a split optical fibre with an output diameter of 25 mm (Schoelly, Denzlingen, Germany). A 15 cm long transparent acrylic rod on the output of the fibre homogenised the light sources. A 3 mm acrylic rod coupled a sample of light within the chamber to a photodiode (OSD15‐5T, Centronic, Croydon, UK) connected in turn to a trans‐impedance amplifier and voltmeter. The calibration procedure of the in‐chamber sensor and the laboratory‐built quantum sensor used for that purpose are described in Taylor et al. (2019). Briefly, this involved the use of a calibrated laboratory‐made light sensor with an optical window diameter equal to that of the inner diameter of the chamber seals and comprising 13 concentrically arranged cosine‐corrected photodiodes embedded across its face. A benefit of this device compared to a conventional single‐point detection quantum flux metre is that it provides an average chamber light intensity as it is known that chamber light distribution can be heterogenous (Hogewoning et al. 2010). A correction was applied for each actinic spectrum to account for the comparatively high reflectivity of the lab‐built sensor, which would have reflected more actinic light onto the chamber walls and back onto itself than a leaf would have done (hence leading to an underestimation of light intensity). This involved comparing irradiance intensity measured by a spectroradiometer (USB2000, Ocean Optics, Duiven, The Netherlands) using a glass fibre with its distal end positioned through a punched leaf or a white polyacrylamide disc (the same material as the optical window of the lab‐built sensor). The irradiance spectrum measured with the glass fibre positioned through the leaf was used to determine the Relative Quantum Flux (RQF) of all irradiance sources used in this study.

2.4. Chlorophyll Fluorescence

Chlorophyll fluorescence was determined using a modulated system by measuring the fluorescence produced by a red (660 nm) measuring beam modulated at approximately 1 kHz (the frequency was adjusted to minimise noise) with an intensity of 0.5 µmol m^{− 2} s^{− 1} (or 1 µmol m^{− 2} s^{− 1} for improved signal to noise ratio when 720 nm or SHADE was used). The output of the optical fibre which delivered the measuring beam to the chamber was filtered with a hot mirror to prevent near‐infra‐red fluorescence produced by the glass‐fibre from entering the leaf chamber. A saturating light pulse (12 000 µmol m⁻² s^{− 1}) was generated using three high power red ‘Phlatlite’ LEDs (Luminus Devices, Sunnyvale, USA) connected to a timer which pulsed the LEDs for 1 s. Three photodiodes (G1736, Hamamatsu, Hamamatsu city, Japan), each filtered with an RG‐9 filter (Schott, Mainz, Germany) and spaced evenly on a PCB of the sensor cluster below the leaf were used to detect fluorescence. A lab‐built phase‐sensitive detector and amplifier recovered and amplified fluorescence signals and these signals were recorded using a data logger (National Instruments, Austin, Texas, USA) with a 10 Hz low‐pass filter on its input to limit bandwidth. All fluorescence parameters were determined and calculated according to Maxwell and Johnson (2000).

The fluorescence yield when all Q_A (the primary stable electron acceptor of PSII) was reduced (i.e. Fm and Fm'; Baker et al. 2007; Maxwell and Johnson 2000) was measured during the application of the saturating light‐pulse. Fm was measured on a dark‐adapted leaf (20 min in darkness) to allow all rapidly reversible chlorophyll fluorescence quenching to relax, while Fm' was measured on an illuminated leaf. The steady state fluorescence yield (Fs) was measured immediately before the application of the saturating light pulse used to measure Fm'. Measurements of Fo and Fo' (the relative chlorophyll fluorescence yield in the absence of any actinic irradiance and when all Q_A is oxidised [Baker et al. 2007; Maxwell and Johnson 2000]) were made either on dark‐adapted leaves (Fo) before the determination of Fm and following the application of a 1 s pulse of far‐red (720 nm) light to oxidise all Q_A, or on light‐adapted leaves (Fo') immediately after stopping actinic irradiance and the application of a 1 s pulse of far‐red light. The measurement of Fo' was made after the measurement of Fm' once CO₂ fixation had recovered to its pre‐saturating light‐pulse rate. From the measurements of Fm, Fm', Fo, Fo', and Fs, the fluorescence parameters qP, Fv'Fm', and Φ_PSII were calculated (Baker et al. 2007; Maxwell and Johnson 2000).

2.5. Φ_PSI Measurements

Φ_PSI was estimated following the calculation of Baker et al. (2007) using changes in 820 nm absorption (∆820 nm). An LED (ELJ‐810‐228B, Roithner Lasertechnik, Vienna, Austria), the output of which was filtered, was used to produce the 820 nm signal. This signal was modulated at 455 kHz and coupled to the chamber by means of a glass fibre. Three silicon photodiodes (BPW 34 FA, Osram, Regensburg, Germany) connected to a selective amplifier containing a 455 kHz ceramic filter (Murata, Kyoto, Japan) were used to recover signal which was recorded by a data recorder. Each ∆820 nm measurement, taken at each light step, involved first determining the 820 nm signal in the light and then in darkness. Obtaining darkness involved interruption of the xenon lamp and/or switching off the LED(s) simultaneously to accommodate instances where both light sources were in use. This was achieved using a mechanical shutter affixed to the output of the xenon lamp which also switched off the LEDs electronically upon closure. Upon reaching steady state signal in darkness, a far‐red pulse of ca. 10 s in duration was applied to oxidise P700 centres during which a 2 ms pulse of saturating irradiance was also applied to oxidise any remaining unoxidised PSI centres (Kingston‐Smith et al. 1999). Operationally, recordings of far‐red absorbance changes were made during the removal of irradiance to make the Fo' measurements and during the subsequent dark period.

2.6. Chlorophyll a/b Ratio

Chlorophyll a/b ratio was determined using the method of Croce et al. (2002). Briefly, an extract was prepared for SUN‐ and SHADE‐grown leaves using buffered 80% acetone. The absorption spectrum of the extract was fitted with that of known absorption spectra for individual pigments to determine chlorophyll a/b ratio.

2.7. Measurement Procedure for the Quantum Yield of Carbon Dioxide Fixation

Quantum yields of CO₂ fixation (Φ_CO2) were determined from slopes of measured assimilation rate versus absorbed PAR, supplied by many different light spectra. Light steps were based on absorbed light intensities of approximately 10, 20, 30, 40, and 50 µmol m^{− 2} s^{− 2}. These light‐limiting irradiance intensities avoid nonlinearity at either the high and the low end (Kok 1948). The slope method provides a more reliable determination of Φ_CO2 than approaches where quantum yields are based on single‐point measurements as it better accounts for respiration in the light and dark which can be difficult to estimate accurately. Since leaf absorptance spectra differed between SUN‐ and SHADE‐grown plants, and because gas exchange measurements on each leaf had to precede the destructive determination of leaf absorptance spectra, the required incident irradiances per actinic irradiance spectrum were estimated based on leaf absorptance spectra of nine representative leaf discs cut from each of three replicate SUN‐ and three replicate SHADE‐grown plants i.e. 3 discs per plant. Two integrating spheres (50 mm diameter reflectance and transmittance spheres; Avantes, Appeldoorn, The Netherlands) were used for this purpose as described in Taylor et al. (2019).

Upon completion of all photosynthetic measurements on each leaf, leaf light absorption was measured using 3 leaf discs cut from the portion of the leaf which was enclosed within the chamber during measurement. The mean of these leaf light absorptance measurements was used to determine Φ_CO2 on an absorbed light basis. Each light‐limited slope was preceded by 20 min of dark‐adaptation, after which dark respiration, estimated from the rate of CO₂ efflux, and F_v /F_m were determined (see above). Light intensity was subsequently increased incrementally, allowing photosynthesis to reach steady state at each intensity (this typically took about 10 min or less per irradiance step) at which point CO₂ and H₂O fluxes were logged over 30 s at a sample rate of 1 Hz. The mean of these records at each light intensity was used to calculate the rate of photosynthesis according to Farquhar et al. (1980). While chlorophyll fluorescence and ∆820 nm were measured at each light step immediately after the CO₂ measurement had been taken, only subsets of these data, measured at the highest absorbed light intensity step used of 50 µmol m^{− 2} s^{− 1}, are presented in the results.

For each of the 17 narrowband spectra described earlier (400–720 nm), Φ_CO2 was determined for three replicate SUN‐ and SHADE‐grown leaves, producing an action spectrum for each leaf type. It took 2 days to complete each action spectrum on a single leaf; at the end of the first day of measurement the chamber was opened slightly to relieve compression by the seals on the leaf which remained fixed in position overnight by the moulding silicon on the lower seal. A further three replicate SUN‐ and SHADE‐grown leaves were used to determine Φ_CO2 when each of the narrowband irradiances was combined with either SUN or SHADE irradiance (SUN irradiance was used for SUN‐grown leaves and SHADE irradiance was used for SHADE‐grown leaves). These spectral combinations involved the halving of intensities of the narrowband irradiances used (i.e. absorbed light intensity steps of 5, 10, 15, 20, and 25 µmol m^{− 2} s^{− 2}), with the resulting halving in absorbed light intensity compensated for by an equal quantity of absorbed light from the background broadband irradiance. This meant that narrowband and broadband irradiances were combined in equal parts (1:1) on an absorbed light basis with identical absorbed light intensity steps used to construct the narrowband light‐limited slopes that is, 10, 20, 30, 40, and 50 µmol m^{− 2} s^{− 2}. Narrowband spectra were applied in random order including when a background irradiance was used.

Mindful of the potential for leaf acclimation over the 2 days it took to complete each set of measurements on a single leaf, each day commenced with a measurement of Φ_CO2 using SUN or SHADE to test for drift of biological (or nonbiological) origin. As differences were negligible, the Φ_CO2 for either SUN and SHADE leaves was taken as the mean of these two measurements. In an attempt to help prevent changes due to acclimation during the measurements, 50 µmol m^{− 2} s^{− 1} of SUN or SHADE on an absorbed light basis (corresponding with the growth spectrum) was also applied between different measurement spectra until gas exchange had stabilised.

2.8. Calculation of Enhancement

The method of (McCree 1972b) was adapted to test for and quantify enhancement. The basis of McCree's approach (adapted here to predict Φ_CO2 instead of assimilation rate) is that Φ_CO2 action spectra (Φ_{CO2 λ}), together with the relative quantum flux (RQF _λ) of the spectrum of interest, can be used to predict Φ_CO2 in the absence of enhancement for that spectrum as described by the equation:

$${\Phi }_{C{O}_{2}predicted}=\sum _{{\lambda }_{min}}^{{\lambda }_{max}}{\Phi }_{C{O}_2(\lambda )}\times RQ{F}_{(\lambda )}$$ (1)

where λ _min and λ _max are the lower and upper wavelength boundaries for the prediction and where RQF is expressed as the fraction of flux per wavelength relative to total flux. An enhancement factor (E) can then be quantified as the ratio of measured to predicted Φ_CO2 for the spectrum of interest, where a value ≤ 1 indicates either no enhancement or negative interference, whereas a value > 1 implies that enhancement did occur:

$$E={\Phi }_{C{O}_{2}measured}⁄{\Phi }_{C{O}_{2}predicted}$$ (2)

The lower and upper wavelength boundaries (λ _min and λ _max) were set to 400 nm and 728 nm, respectively, to assess enhancement in SUN and SHADE‐grown leaves in corresponding SUN and SHADE actinic spectra as well as in 1:1 combinations (on an absorbed light basis) of SUN and SHADE with each of the 17 narrowband spectra. The upper wavelength limit was chosen based on leaf light absorption not being distinguishable from zero at a wavelength of 728 nm in SUN and SHADE‐grown leaves (Figure 2), and no photosynthesis was measured in irradiance with a peak wavelength at 740 nm. The procedure to calculate Φ_{CO2 predicted} was as follows: first the Φ_CO2 data, measured using each of the 17 narrowband irradiances, was linearly interpolated to produce a nanometre‐by‐nanometre action spectrum for photosynthetic light‐use efficiency, one for SUN leaves and one for SHADE leaves. Action spectra were then multiplied by the corresponding nanometre‐by‐nanometre relative quantum flux of the spectrum of interest. These products were summed to produce a predicted Φ_CO2 value for that spectrum. The enhancement factors for each irradiance (SUN or SHADE, or SUN and SHADE plus a narrowband spectrum) were calculated using Equation 2.

Figure 2.

Leaf absorptance of SUN‐grown leaves (solid line) and SHADE‐grown leaves (dashed line) (n = 3). Leaf absorptance was measured using two integrating spheres (one to measure leaf reflectance and one to measure leaf transmittance), a USB 2000 spectrometer, and LED light sources (Taylor et al. 2019).

2.9. Experimental Design, Data Analysis and Statistics

Plants were raised in a staggered sowing schedule over a period of 2 months. Measurements commenced on three replicate SUN leaves using each of the 17 narrowband irradiances and, separately, the SUN spectrum. The next set of measurements on SUN leaves involved the combination of each of the 17 narrowband irradiances with the SUN broadband spectrum in a 1:1 ratio on an absorbed light basis, but SUN alone was also applied separately as with the narrowband measurements. SHADE leaves were subsequently measured in the same way as SUN leaves, with the only difference being that the broadband spectrum was SHADE in place of SUN.

Φ_CO2 was determined from the slope of a linear regression of the response of CO₂ fixation to absorbed light intensity. The quality of the regression analysis was assessed using the coefficient of determination (R ²) combined with visual inspection of the scatter plots to check for any deviations from a linear trend. The means of measured and predicted Φ_CO2 were compared using one‐way analysis of variance tests to test for statistical significance, with significant tests followed by Tukey's HSD for post‐hoc multiple comparisons (p = 0.05). These tests were performed using IBM SPSS Statistics for Windows Version 30.0 (IBM Corp., Armonk, NY, USA).

3. Results

The two growth spectra (SUN and SHADE) were chosen to produce leaves with different acclimation histories. The SHADE growth spectrum produced two of the classical features associated with growth in the shade: plants were significantly more etiolated than SUN‐grown plants and the chlorophyll a/b ratio in leaves of SHADE‐grown plants was 2.89 compared with that of 3.08 in SUN‐grown plants. Leaf absorptance tended to be slightly lower in SHADE‐grown leaves than in SUN‐grown leaves, with the largest difference at approximately 550 nm. Leaf absorptance reduced slightly faster in SHADE than in SUN‐grown leaves at wavelengths greater than around 680 nm, and could not be resolved from zero in both leaf types at about 728 nm (Figure 2).

3.1. Response of the Quantum Yield of Carbon Dioxide (Φ_CO2) to Narrowband Irradiances Alone

The response of CO₂ fixation to light was highly linear between 10 and 50 µmol m^{− 2} s^{− 1} in the PAR spectral region. The coefficient of variation (R ²) was high (≥ 0.99) for all measurements apart from in SHADE‐grown leaves under 720 nm where mean R² was somewhat lower (0.98). Quantum yields for CO₂ fixation (Φ_CO2) on an absorbed light basis for SUN‐grown and SHADE‐grown leaves subjected to narrowband irradiances were greatest in the red region with a shoulder at 406 nm to 440 nm and a trough at 500 nm (Figure 3). A sharp drop in Φ_CO2 occurred above 680 nm. Φ_CO2 was generally comparable for SUN‐ and SHADE‐grown leaves in the blue and blue‐green region (i.e., ≤ 520 nm). However, the effect of growth spectrum on wavelength dependence of Φ_CO2 was marked at wavelengths of 540 nm and greater, with Φ_CO2 of SHADE‐grown leaves lower than SUN‐grown leaves between 540 nm and 680 nm. At wavelengths greater than 700 nm Φ_CO2 was greater in SHADE‐grown leaves. Maximum mean Φ_CO2 was 0.090 and 0.085 for SUN‐grown and SHADE‐grown leaves, respectively. The minimum Φ_CO2 within the PAR region was 0.064 in SUN‐grown leaves and 0.063 in SHADE‐grown leaves, occurring at 500 nm.

Figure 3.

Wavelength dependence of Φ_CO2 on an absorbed light basis for SUN‐grown (open circles) and SHADE‐grown (closed circles) tomato leaves using 17 narrowband irradiances. Each narrowband irradiance was applied individually. The wavelength (x‐axis) is the peak wavelength of each narrowband spectrum, which typically had a spectral width at half‐maximum intensity of 10–15 nm. The light sources used were LEDs combined with bandpass interference filters. Φ_CO2 was measured as the slope of the relationship between the narrowband irradiance and gross CO₂ fixation rate. The nominal narrowband irradiance intensities used corresponded with 10, 20, 30, 40, and 50 μmol m^{− 2} s^{− 1} of absorbed light (400–728 nm) with the actual absorbed irradiance being calculated after the measurement of the light response curve. Error bars are ±SE of the mean and are shown where the bars are larger than the data points (n = 3).

3.2. The Effect on Φ_CO2 of Combining Broadband Spectra Irradiances and Narrowband Irradiances

Φ_CO2 for SUN‐grown leaves subjected to SUN, and SHADE‐grown leaves subjected to SHADE, are shown in Figure 4 together with their predicted Φ_CO2. In SUN‐grown leaves measured Φ_CO2 was 0.069 and predicted Φ_CO2 was 0.073 (with a corresponding enhancement value (E‐value) of 0.94) but this difference was not significant (p = 0.14). However, the difference between measured Φ_CO2 in SHADE‐grown leaves (0.075) and the predicted Φ_CO2 (0.061) was significant (p < 0.001) and yielded an E value of 1.23.

Figure 4.

Measured (open bars) and predicted Φ_CO2 (assuming no enhancement; shaded bars) for SUN‐ and SHADE‐grown leaves in SUN and SHADE light. Φ_CO2 was measured as the slope of the relationship between the narrowband irradiance and gross CO₂ fixation rate. The nominal SUN and SHADE irradiance intensities used corresponded with 10, 20, 30, 40, and 50 μmol m^‐2 s^‐1 of absorbed light (400–728 nm) with the actual absorbed irradiance being calculated after the measurement of the light response curve. A xenon arc‐lamp was used to produce the SUN spectrum (Newport Instruments, USA), filtered by a Calflex NIR filter (Balzers, Lichtenstein). For the SHADE spectrum a quartz‐halogen lamp was filtered. Error bars are ±SE of the mean (n = 3).

In general, the trends in wavelength dependency of Φ_CO2 were preserved when a background of SUN or SHADE was used (Figure 5, row A). For example, the trough at 500 nm and the peak in the red region were still evident in both leaf types. However, as expected, the variation in Φ_CO2 values were considerably reduced as a result of the moderating effect of the broadband component. The maximum Φ_CO2 with background SUN or SHADE occurred when 680 nm irradiance was used irrespective of whether leaves were SUN‐ or SHADE‐grown; in SUN‐grown leaves this value was 0.083 with SUN as background and in SHADE‐grown leaves this value was 0.089 with SHADE as background. The corresponding minima Φ_CO2 for SUN‐ and SHADE‐grown leaves in the 400–700 nm range were 0.068 and 0.074, respectively, when a foreground narrowband irradiance of 500 nm was used, coinciding with the minima observed at 500 nm when narrowband spectra were used alone. SHADE was used more efficiently by SHADE‐grown leaves than SUN by SUN‐grown leaves (0.075 compared with 0.069).

Figure 5.

(Row A) Φ_CO2 obtained when narrowband irradiances were applied with SUN (left panes) or SHADE (right panes) as background irradiance in a 1:1 ratio on an absorbed light basis (closed circles). Wavelength dependence data from Figure 3 are shown for comparison (open circles, dashed lines). Dotted lines represent mean Φ_CO2 when the background irradiance (SUN or SHADE) was applied alone. Row B: Predicted Φ_CO2 for each narrowband irradiance with SUN or SHADE as background irradiance, assuming no enhancement (open circles). Predictions were made using Equation 1 which was adapted from McCree (1972b). Measured values from Row A are shown again for comparison (closed circles, dashed lines). Row C: Enhancement factor (E) calculated using Equation 2 as the ratio of measured:predicted Φ_CO2 using data shown in Row B. Error bars are ±SE of the mean and are shown where the bars are larger than the data points (n = 3).

The predicted Φ_CO2 (assuming no enhancement, thus using a linear combination of values) is shown in Figure 5 (row B) with the measured Φ_CO2 from Figure 5 (row A) shown again for reference. From these results the three greatest E values in SUN‐grown leaves with the SUN background were (in sequence from 680 nm to 720 nm foreground irradiance, with p values for the comparison of corresponding measured and predicted values in brackets): 1.047 (p = 0.14), 1.076 (p < 0.001), and 1.46 (p < 0.001). Mean E for the same leaves presented with SUN background and foreground irradiances of < 680 nm was 1.011 (SE = 0.011) indicating that enhancement was negligible at foreground wavelengths < 680 nm. In SHADE‐grown leaves with the SHADE background E values were substantially greater than in SUN‐grown leaves across the range of foreground irradiances used. As with SUN‐grown leaves, the greatest E value in SHADE‐grown leaves occurred with 720 nm foreground irradiance although this value was far greater (1.76), indicating considerable enhancement of 76%. E values in SHADE‐grown leaves at all other supplementary wavelengths were between 1.17 and 1.27. Comparisons of measured and predicted Φ_CO2 for SHADE leaves were significant for all 17 actinic spectra (p < 0.001).

3.3. The Effect of Narrowband Irradiances of Different Wavelengths on the Operation of PSI and PSII

The results of selected PSII parameters, Φ_PSI, and photosystem balance (Φ_PSII/(Φ_PSI + Φ_PSII) are presented in Figure 6. We deal first with leaves subjected to narrowband irradiance alone. Minima in Φ_PSII occurred at 480 nm, 560 nm, and 660 nm in both leaf types (Figure 6, row A). Outside these troughs, Φ_PSII in SUN‐grown leaves was high (ca. 0.75–0.8) but greatest at 720 nm (> 0.8). Though the trend of Φ_PSII in SHADE‐grown leaves exposed to narrowband irradiances alone was similar to that of SUN‐grown leaves, differences were amplified such that maxima were greater than, and the minima less than, in SUN leaves. For example, the Φ_PSII minima in SUN leaves (corresponding values for SHADE‐grown leaves in brackets) at 480, 560 and 660 nm were, respectively, 0.69 (0.64), 0.74 (0.69), and 0.72 (0.70) whereas the maximum at 720 nm was 0.81 (0.83). The sharp peak in Φ_PSII in SHADE‐grown leaves at 520 nm is especially notable compared with the shoulder in this region in SUN‐grown leaves. It is notable that Φ_PSI closely mirrored the response of Φ_PSII in both leaf types; when Φ_PSII was high, Φ_PSI was low and vice versa. For example, Φ_PSI maxima coincided with Φ_PSII minima at 480, 560 and 660 nm whereas Φ_PSI minima coincided with the Φ_PSI maximum at 720 nm and also locally at 400, 520, and 700 nm, when Φ_PSII was high.

Figure 6.

(Row A) The light‐use efficiency for PSII electron transport (Φ_PSII), (Row B) The light‐use efficiency for PSI electron transport PSI (Φ_PSI), (Row C) The balance of photosystem efficiency taken as the ratio of Φ_PSII/(Φ_PSI + Φ_PSII), (Row D) the photosynthetic efficiency factor (qP), a measure of the effect of PSII trap‐closure upon Φ_PSII, and (Row E) maximum PSII efficiency in the light (i.e., the value of Φ_PSII when qP = 1; Fv’/Fm’), measured at the highest absorbed light intensity of ca. 50 µmol m^{− 2} s^{− 1} for SUN‐ (left panes) and SHADE‐grown (right panes) leaves. Measurements were taken using narrowband irradiances only (open circles) or with SUN or SHADE background irradiance in a 1:1 ratio on an absorbed light basis (400–728 nm). Dotted lines represent values for the values measured with the corresponding SUN or SHADE spectra alone. Error bars are ±SE of the mean and are shown where the bars are larger than the data points (n = 3).

The wavelength dependence of photosystem balance, taken as the ratio of Φ_PSII to the sum of Φ_PSII and Φ_PSI (Figure 6, row C), generally ranged between ca. 0.4–0.45 in both leaf types and tended to follow the Φ_PSII trend (since for most wavelengths changes in Φ_PSII were comparatively greater than that of Φ_PSI). Note that for our leaves typical maxima for Φ_PSII were about 0.8 and for Φ_PSI it was 0.98, giving a photosystem balance of 0.45. However, at > 700 nm the marked loss of Φ_PSI had a dramatic impact on the relative efficiencies of the photosystems, with Φ_PSII/(Φ_PSI + Φ_PSII) ratios of 0.87 and 0.84 in SUN‐ and SHADE‐grown leaves, respectively, at 720 nm. The qP parameter (Figure 6, row D) quantifies the effect on Φ_PSII of the closure of PSII reaction centres as a result of Q_A reduction (Baker et al. 2007) and is qualitatively correlated with Q_A reduction. The variation of qP showed striking similarity to that of Φ_PSII in both leaf types, with loss of qP coinciding with, and contributing to, Φ_PSII minima where PSII overexcitation occurred. Φ_PSII is the product of qP and Fv'/Fm', where Fv'/Fm' is the maximum efficiency of PSII in the light (i.e. if or when qP = 1). Fv'/Fm' showed little response to wavelength, apart from increases at 700–720 nm. Although values in SUN‐grown leaves tended to be slightly less than 0.8, in SHADE‐grown leaves values were slightly greater than this, supporting the conclusion that changes in Φ_PSII are strongly dependent on qP.

3.4. The Effect of Combining Broadband Spectra With Narrowband Spectra on the Operation of PSI and PSII

Broadband spectra had a considerable moderating effect on Φ_PSII, qP, and the photosystem excitation balance (Figure 6, row A–C). Φ_PSII varied within a tight range between 0.77 and 0.81 in SUN‐grown leaves and 0.76 and 0.79 in SHADE‐grown leaves. Excitation balance was about 0.45 in both leaf types apart from at 720 nm. Troughs in these parameters observed for narrowband irradiance alone at 480 nm and 680 nm were preserved but only to a reduced extent. Unlike Φ_PSII and qP, the response of Φ_PSI to a combination of narrowband and broadband (SUN and SHADE) irradiances was little different to that obtained under narrowband irradiances alone (Figure 6, row B). The main difference between the leaf types was in the spectral response of Fv'/Fm' (Figure 6, row E). Whereas this parameter was largely uniformly higher by about 0.03 at most wavelengths in SUN‐grown leaves in the presence of the SUN background (the exceptions being at 550 nm and over 700 nm where this increase did not occur), SHADE‐grown leaves with the SHADE background showed no difference apart from at 700 and 720 nm.

Since Φ_PSII is the product of the proportion of open reaction centres (qP) and the maximum efficiency of Φ_PSII in the light (Fv'/Fm'; see above), the values of these underpin the spectral response of Φ_PSII. Of qP and Fv'/Fm' it is variation in qP that most greatly effects Φ_PSII; the relationships between qP and Φ_PSII (Figure 7) are nearly linear. Figure 7 also shows that, in both leaf types, greater variation in qP occurred when only narrowband irradiances were used, and comparatively more variation was observed in SHADE‐grown leaves than in SUN‐grown leaves. Only in SUN‐grown leaves does the small step‐change in Fv'/Fm' described earlier (the increases occurring at 700–720 nm) lead to the separation of two discernible relationships, depending on whether narrowband irradiance was used alone (open points) or in combination with the SUN spectrum (closed points).

Figure 7.

Relationship between qP and PSII for SUN‐ (A) and SHADE‐grown (B) leaves, measured at the highest absorbed light intensity of ca. 50 µmol m^{− 2} s^{− 1}. Measurements were taken using narrowband irradiances only (open circles) or with either SUN or SHADE background irradiance in a 1:1 ratio on an absorbed light basis (400–728 nm). Φ_PSII is the light‐use efficiency for electron transport by PSII and qP is the photosynthetic efficiency factor (qP), a measure of the effect of PSII trap‐closure upon Φ_PSII.

3.5. State Transitions in Response to Narrowband Irradiances or Combined Narrow‐ and Broadband Spectrum Irradiances

State transitions were quantified using the ratio of light‐adapted maximum fluorescence (Fm') to dark‐adapted maximum fluorescence (Fm) (Allen 1992; Figure 8). In SUN‐grown leaves the spectral response of Fm'/Fm to narrowband irradiance alone was similar to that of Φ_PSII and qP. For example, the trough in Φ_PSII at 480–500 nm and local maxima occurring at 520–540 nm, 640 nm, and 700–720 nm are paralleled by the response of Fm'/Fm. In response to the overexcitation of PSII at 480–500 nm, state 2 was induced as shown by the lower Fm'/Fm ratio of 0.9. At 520 nm and 640 nm, when PSII excitation pressure was comparatively low, Fm'/Fm was ≈1, indicating state 1. Unexpectedly, Fm'/Fm was greater than 1 at 700 and 720 nm. Also, in SUN‐grown leaves the combination of the SUN spectrum with the narrowband irradiances tended to moderate Fm'/Fm values around 1 (with the exception of 700 and 720 nm), indicating the predominance of state 1 (i.e. no migration of LHCII to PSI). On the other hand, SHADE‐grown leaves subjected to either the narrowband irradiance alone or in combination with the SHADE spectrum tended to have Fm'/Fm values of less than 1, indicating the predominance of state 2, irrespective of spectrum. An exception to the latter occurred when using 720 nm narrowband irradiance alone where Fm'/Fm was greater than 1.

Figure 8.

Fm’/Fm for SUN‐ (A) and SHADE‐grown (B) leaves, measured at the highest absorbed light intensity of ca. 50 µmol m^{− 2} s^{− 1}. Measurements were taken using narrowband irradiances only (open circles) or with SUN or SHADE background irradiance in a 1:1 ratio on an absorbed light basis (400–728 nm). Fm is the maximum relative quantum yield for chlorophyll fluorescence measured in the presence of a saturating light pulse which saturates the value of Fm; Fm’ is similar except that it is measured in the presence of an actinic irradiance that provokes changes in Fm’ due to the action of non‐photochemical quenching. Dotted lines represent values for the corresponding SUN or SHADE spectra. Error bars are ±SE of the mean and are shown where the bars are larger than the data points (n = 3).

Since changes in Fm'/Fm are proportional to changes in cross‐sectional area of PSII, which can impact the excitation pressure on PSII, the relationship between qP and Fm'/Fm was explored (Figure 9). For SUN‐ and SHADE‐grown leaves subjected to narrowband irradiance alone, the general trend was an increase in qP as Fm'/Fm increased. When Fm'/Fm was high (> 1), qP was also high (0.95–1), occurring at PSI irradiances when a transition to state 1 increased the cross‐sectional area of PSII. The lowest Fm'/Fm values were about 0.9 in SUN‐grown leaves and 0.85 in SHADE‐grown leaves (state 2). At around these Fm'/Fm values qP decreased sharply to a minimum of about 0.87 in SUN‐grown leaves and 0.8 in SHADE‐grown leaves. In SUN‐ and SHADE‐grown leaves subjected to narrowband plus their corresponding broadband spectra, the data points were tightly clustered although in SHADE‐grown leaves slight decreases in qP were observed when Fm'/Fm was at its lowest.

Figure 9.

Relationship between Fm’/Fm and qP for SUN‐ (A) and SHADE‐grown (B) leaves, measured at the highest absorbed light intensity (400–728 nm) of ca. 50 µmol m^‐2 s^‐1. Measurements were taken using narrowband irradiances only (open circles) or with SUN or SHADE background irradiance in a 1:1 ratio on an absorbed light basis (400–728 nm). Error bars are ±SE of the mean and are shown where the bars are larger than the data points (n = 3).

4. Discussion

4.1. Spectral Quantum Yield Results

The spectral quantum yield results for CO₂ fixation in tomato (Φ_CO2) measured under narrowband irradiance alone (Figure 5) agree closely with those of Hogewoning et al. (2012) measured on cucumber, which were also measured on an absorbed light basis. A key difference, however, is that in cucumber lower yields in the 400–480 nm spectral region were observed, which formed part of a broader loss of Φ_CO2 extending from 400 nm to 560 nm (600 nm in the case of their shade light‐grown plants). Hogewoning et al. (2012) attributed this loss to light absorption by non‐photosynthetic pigments at shorter wavelengths and the inefficiency of carotenoid to chlorophyll excitation transfer (e.g. Croce et al. 2001; de Weerd et al. 2003; Wientjes et al. 2011). It has been estimated that carotenoids reduce blue light use efficiency by the photosystems by about 30% (Laisk et al. 2014). The relatively higher Φ_CO2 which we observed in the 400–480 nm spectral region may be due to the possible relative absence of non‐photosynthetic pigments in our tomato cultivar (cf. Hogewoning et al. 2012). In any case the loss of photosynthetic light‐use efficiency in the 480–600 nm region is shared by tomato and cucumber, and was also observed by Inada (1976) and McCree (1972a) for a wider range of species.

The greatest Φ_CO2 obtained using broadband irradiance in the present study was 0.089, obtained using SHADE‐grown leaves presented with SHADE enriched with 680 nm irradiance. This value compares favourably with 0.092 for pea leaves (Evans 1987) and a mean of 0.093 for a variety of 11 C3 species (Long et al. 1993). In both those studies, a quartz‐iodine lamp (which has a spectrum close to that of a 3200 K black‐body emitter) was used for actinic light. The output from such a lamp comprises a significant fraction of red but also less efficiently used far‐red wavelengths, so in some ways is comparable to the SHADE spectrum enriched with 680 nm irradiance. It is likely that significant enhancement also occurred in those studies because, despite the less efficiently used wavelengths, their Φ_CO2 values are still comparable to the maximum Φ_CO2 of 0.093 reported by Hogewoning et al. (2012) for artificial sunlight‐grown cucumber leaves subjected to 620 nm narrowband irradiance and the maximum Φ_CO2 obtained in the present study of 0.090 when using 680 narrowband irradiance. Singsaas et al. (2001) concluded that the maximum quantum yield deviates little across diverse taxa from the mean maximum value of 0.093 reported by Long (1993) and that variation in this value is likely due to measurement error. The use of 53 unique spectra in the present study allowed us to rigorously test the effect of actinic spectrum on this putative maximum value. The maximum Φ_CO2 value obtained from our extensive analysis provides support for the conclusion of Singsaas et al. (2001) and also suggests that this supposed upper limit is not exceeded even in instances where significant enhancement occurs.

4.2. The Presence of the Enhancement Effect

We show that spectral enhancement of photosynthetic efficiency in tomato leaves operating in permanent foliar shade can be significant but is absent in leaves that are grown in (low intensity) sunlight (Figure 4): An enhancement of 23% occurred in SHADE‐grown leaves presented with the SHADE irradiance. This enhancement is considerably greater than the maximum of 7% for four different types of “white” light (McCree 1972b) but comparable to the 21% enhancement observed in cucumber leaves grown and measured using the same artificial SHADE irradiance which we used (Hogewoning et al. 2012). Given that the SHADE spectrum comprises considerably more far‐red (about 28% of total incident photon flux at wavelengths beyond 700 nm in the band 700–728 nm compared with a corresponding figure of 10% for the SUN spectrum), the far‐red in the SHADE spectrum is an obvious candidate for further investigation, especially given its role in classical far‐red enhancement. The greatest enhancement in SUN and SHADE leaves occurred when SUN and SHADE spectra were enriched with 720 nm (46% and 76%, respectively), which is consistent with reports of far‐red light enhancement in broad‐band ‘white’ light. For example, apparent Φ_CO2 increased by up to 41% in lettuce when far‐red light (700–770 nm) was added to warm white LED light (Zhen and van Iersel 2017). In selected species of green algae, Dring and Lüning (1985) calculated an enhancement of 18% when using an irradiance from a quartz‐iodine lamp that was rich in far‐red but low in blue, and calculated that there was no enhancement in irradiance spectra lacking far‐red. Turning attention to the absence of enhancement in our SUN spectrum, we cannot resolve the 10% enhancement, also for an artificial daylight spectrum, calculated by Hogewoning et al. (2012). This difference may be at least partly attributable to the different artificial daylight growth and actinic spectra used in our study.

An unexpected result in our study was the strong enhancement observed when the SHADE spectrum was combined with the 720 nm irradiance. The 720 nm irradiance strongly overexcites PSI but the SHADE spectrum, itself being rich in far‐red, can also be considered PSI light. Therefore, the combination of the two irradiances does not intuitively lead to the expectation of enhancement. However, in our SHADE spectrum shorter wavelengths are present that drive PSII photochemistry more strongly than the 720 nm irradiance which itself drives PSII only very weakly; the complementary action of the shorter ‘PSII’ wavelengths with the 720 nm ‘PSI’ irradiance increases photosynthetic efficiency hence the observed enhancement. It is this complementary action which underpins enhancement and therefore it is not one irradiance which enhances another so much as both irradiances enhancing the other. While the full explanation of this phenomenon must wait until the discussion of the regulation of PSII, these findings demonstrate that, while the use of PSI and PSII narrowband irradiances simplifies an understanding of the enhancement phenomenon, predicting whether enhancement may occur in broadband spectrum demands a holistic view of the component spectra within the broadband spectrum, itself often being a mixture of PSI and PSII irradiances in various proportions.

Apart from differential excitation of PSI and PSII, the response of photosynthesis to mixtures of irradiance could also be influenced by the different penetration into the leaf of the different actinic wavelengths used. Light absorption by a leaf is a complex process, depending on both the absorption and the scattering of light within the leaf (Merzlyak et al. 2009). The absorption of light is a prerequisite for photosynthesis, though not all light that is absorbed by a leaf is absorbed by photosynthetic pigments (Cerovic et al. 2002; Hogewoning et al. 2012). The use of photosynthetically active light would be expected to lead to the formation of lightintensity gradients within the leaf (e.g. Cui et al. 1991; Vogelmann and Han 2000) with less strongly absorbed wavelengths (such as those in the green or far‐red parts of the spectrum) resulting in shallower intensity gradients than more strongly absorbed wavelengths. These gradients could influence the development of enhancement by affecting the ratio of PSI and PSII wavelengths as the actinic irradiance penetrates the leaf.

4.3. Responses of PSI and PSII; Φ_PSI, Φ_PSII, qP and Fv'/Fm'

The differential penetration of light into leaves is also an important consideration when interpreting optically‐measured chlorophyll fluorescence and PSI results. In our configuration, the measuring beam used to measure P700 oxidation must penetrate the leaf to reach the detector, and it is also strongly scattered by the leaf (Harbinson and Woodward 1987). This implies that the P700 oxidation measurement samples throughout the depth of the leaf. In contrast, the red excitation beam which we applied from above the leaf to measure chlorophyll fluorescence is strongly absorbed by photosynthetic pigments and so the strongest fluorescence signals will tend to be produced by the uppermost part of the leaf where absorption of excitation light is strongest. These differences of how light interacts with leaves and how different measurement techniques sample the leaf is an issue for the quantitative analysis of leaf photosynthesis (Kingston‐Smith et al. 1997; Harbinson 2018). We tested for potential artefacts of light gradients within our leaves on measured Φ_PSII by comparing the irradiance response of Φ_PSII determined using either 660 nm or 560 nm excitation (Figure S2A). The values for Φ_PSII based on 660 nm or 560 nm excitation are very similar (contrast this result with Figure 3 of Kingston‐Smith et al. [1997]). They are also both very similar to Φ_PSI, a measurement made using a weakly absorbed far‐red measuring beam (Figure S2B). This suggests that there were no strong measurement gradients within the tomato leaves we worked on and that parameters associated with PSI and PSII can be compared. Furthermore, variation in the parameters we measured, while being wavelength dependent, did not generally show the same wavelength dependency as leaf light absorption, the main exception to this being the decrease in leaf‐light absorption in the near infra‐red region (ca. 690 nm and above) correlating with an excitation imbalance between PSI and PSII (Figure 6, row C). This general lack of correlation between photosynthetic parameters and the wavelength profile of leaf light absorption also implies that changes in parameters are not driven primarily or simply by changes in light penetration.

While we believe there are no major effects on our analysis of enhancement and our comparison of PSI and PSII parameters, there are still some more minor effects that could be due to problems with the measurement of PSI and PSII of the kind we have summarised above. For example, the SUN‐grown leaves of tomato (Figure 6, row A), in response to the use of narrowband irradiance alone, show a relatively high Φ_PSII in the 400–450 nm region, and relatively high Φ_PSII around 520 nm. Φ_PSI (Figure 6, row B) in the same regions shows decreases, which is broadly to be expected. If PSI is being overexcited relative to PSII then it would be expected that ΦPSII at these low irradiances would have a value of 0.8 or higher rather than the value of 0.76 that was measured. One possible explanation for this apparent simultaneous limitation by both PSII and PSI could be due to measurement gradients within the leaf.

The results we obtained for the spectral dependency of Φ_PSII, qP, Fv'/Fm', and Φ_PSI in response to a narrowband irradiance are generally strikingly similar to those obtained by Hogewoning et al. (2012) in cucumber. The similarity between the spectral responses of Φ_PSII for SHADE‐grown leaves of tomato and cucumber is more conspicuous because of the stronger variation of Φ_PSII in SHADE‐grown leaves. Our results for the dependency of Φ_PSII in tomato to narrowband irradiance also agree with those of Mattila et al. (2020), who used an ingenious technique using changes in Fo to monitor PSI/PSII balance in tobacco and HPLC to measure plastoquinone redox state. The main difference between their Fo‐derived results and ours is that in their case the wavelengths of their overexcitation‐of‐PSII maxima (460–500 nm, 560–570 nm, and 650–660 nm) show similar degrees of overexcitation whereas our results show the greatest overexcitation in the blue region, followed by the red and then the green. The results they obtained for the wavelength dependency of plastoquinol redox state parallel our results for Φ_PSII, Φ_PSI, and (Φ_PSII/Φ_PSII + Φ_PSI) but their data (their Figure 7B) shows stronger fluctuations with wavelength than does ours.

The symmetry of the wavelength dependence of our Φ_PSII and Φ_PSI results (i.e. when Φ_PSII was high, Φ_PSI was low, and vice versa) is to be expected given the underlying symmetry of the wavelength dependance of photosystem absorptance, as reported by Laisk et al. (2014). The extent to which a given spectrum may imbalance the photosystems will depend on the relative amounts of PSII and PSI, which is a plastic trait influenced strongly by growth spectrum (Chow et al. 1990; Hogewoning et al. 2012). The SHADE‐grown leaves, likely having more PSII LHCs to compensate for the PSI‐overexciting growth light, were predisposed to photosystem imbalance in most of the narrowband irradiances given that PAR wavelengths generally over‐excite PSII.

4.4. Responses of PSI and PSII; State Transitions, and the Interaction With qP

The susceptibility of SHADE leaves to PSII overexcitation is further reflected by the tendency of SHADE leaves to be configured in state 2, as evidenced by the comparatively left‐shifted spread of Fm'/Fm points in Figure 9. It is apparent that state transitions in both leaf types were not sufficient to compensate for strong imbalances as demonstrated by the troughs in Φ_PSII/(Φ_PSII + Φ_PSI) at strong PSII irradiances such as 480 and 660 nm. Furthermore, in response to strong PSII‐overexcitation, qP fell sharply once Fm'/Fm showed no further decreases (Figure 9), revealing the limited scope of state transitions to arrest sharp decreases in qP beyond, in this instance, a transition to state 2. These findings are consistent with the findings of Taylor et al. (2019) who observed that state transitions mitigate, but do not eliminate, Φ_CO2 losses caused by strong photosystem imbalances. Nonetheless, the influence of state transitions on Φ_CO2 is of significance to this study as state transitions would have served to ameliorate the response of Φ_CO2 to irradiances which induced photosystem imbalance.

In both tomato (this study) and cucumber (Hogewoning et al. 2012) changes in Φ_PSII in response to narrowband irradiance alone were due to changes in qP, one of the two components of Φ_PSII (the other being Fv'/Fm'). In cucumber (Hogewoning et al. 2012), variation in Φ_PSII was independent of Fv'/Fm' while for our work on tomato changes in Fv'/Fm' were largely independent of the applied narrowband wavelength (Figure 6, row E). SUN‐grown leaves illuminated with narrowband light alone showed the greatest variation in Fv'/Fm' with weak decreases in Fv'/Fm' at those wavelengths where qP was low (around 475 nm and 650 nm). This could be due to only these leaves showing a marked wavelength specific variation in Fm/Fm'. If LHCII attaches to PSI it would result in an increase in PSI fluorescence which would result in a lowering of Fv'/Fm'. This is because a fixed fractional decrease in PSII fluorescence combined with a fixed fractional increase in PSI fluorescence will have a relatively greater effect on Fo' than Fm' (and Fv'=Fm'‐Fo'); the fraction of PSI fluorescence in the far‐red spectral region where we measured chlorophyll fluorescence is about 40% of the Fo' level but only about 10% of the Fm' level (Bos et al. 2023). With a 40% fraction of PSI fluorescence and a Φ_PSII of 0.8, a decrease of Fm'/Fm from 1.0 to 0.9 would result in a decrease in the value of Fv'/Fm' by 0.02. Significantly, in the SUN‐grown leaves in both narrowband irradiances and narrowband irradiances plus SUN, and SHADE‐grown leaves in only narrowband irradiance, Φ_PSII and qP were high at 700 nm and 720 nm. In these treatments Fm'/Fm was also conspicuously high, and Fv'/Fm' was relatively high, consistent with the idea that LHCII migration can produce small changes in Fv'/Fm' as was also shown in Taylor et al. (2019). The increase of Fm'/Fm above 1.0 that was sometimes observed in the presence of far‐red light (e.g. Figures 8 and 9) was unexpected given that dark‐adaptation (or far‐red light) is believed to lead to the thylakoids being in state 1 (i.e. all LHCII is associated with PSII; Box 1 of Haldrup et al. 2001). An increase in Fm' relative to Fm suggests that a dark‐adapted leaf may not have been fully in state 1. The detachment of LHCII from PSII and its association with PSI is believed to be driven by phosphorylation of LHCII (Allen 1992). Greater phosphorylation of LHCII has been observed in the dark‐adapted leaves of spinach than in far‐red illuminated leaves, accompanied by fewer PSI‐LHCI‐LHCII supercomplexes in the far‐red illuminated leaves than in the dark‐adapted leaves (Wood et al. 2019).

It is important to note that state transitions, like qE (Murchie and Harbinson 2014), are a component of non‐photochemical quenching (npq; e.g., Quick and Stitt 1989) and need to be reliably distinguished from qE. In this study, we believe we ruled out any contribution from qE for the following reasons: Firstly, the low irradiances we used were well within the light‐limited region of the light‐response curve where qE is not expected to engage. Secondly, in an analysis of changes in Fm' in relation to state transitions in leaves of tomato grown identically to the plants used in this study, the changes in Fm' were consistent with state transition rather than qE (Taylor et al. 2019). Finally, the high values of Fv'/Fm' (which decreases as qE increases) in our leaves (Figure 6, row E) would suggest little or no qE was present.

4.5. Combining Things; the Enhancement of Shade‐Light by PAR and 720 nm Narrowband Irradiance

The comparatively skewed PSI/PSII absorbance towards PSII in SHADE leaves means that adding narrowband light alone to SHADE‐grown leaves will provoke a greater loss of Φ_PSII (and fall of qP) at most wavelengths below 700 nm, compared to SUN‐grown leaves (Figure 6, row A). If the narrowband light is added in the presence of SHADE light this decrease is much reduced (Figure 6, row A, column SHADE‐grown), implying that the extra far‐red in the SHADE spectrum to some extent re‐balances the imbalance between PSI and PSII. SHADE‐light alone has the highest Φ_PSII/(Φ_PSI + Φ_PSII) of any irradiance below around 680 nm implying it tends to over‐excite PSI. When applying 700 nm or 720 nm alone, the Fm'/Fm is greatest in SHADE leaves (Figure 8)–even higher than 1–implying that even in the dark‐adapted state some LHCII is attached to PSI. These far‐red wavelengths on their own, however, even with the PSII antenna maximised (a high Fm'/Fm), result in only a low Φ_CO2 (Figure 5, row A). Adding the broad‐spectrum SHADE light (which contains shorter wavelengths) to the 720 nm narrowband irradiance will provide not only PSII wavelengths missing in the 720 nm narrowband irradiance but will also activate a state transition, as shown by the lowered Fm'/Fm at 720 nm in Figure 8 for narrowband + SHADE. A similar response is seen when SUN light is added to a SUN‐grown leaf along with 720 nm narrowband irradiance but this effect is not so strong. This state transition will increase the absorption of (short‐wavelength) light by PSII, and so will increase the enhancement effect (Figure 5, row C) seen when broad‐spectrum SHADE light is added to a 720 nm narrowband irradiance.

It can be concluded that enhancement was significant under many of the spectral combinations presented here. A commonality amongst instances where enhancement was found to occur is a spectrum rich in far‐red light such as the SHADE spectrum, or when supplementary 720 nm irradiance was combined with the SUN or SHADE spectra. Though unremarkable at the level of leaf assimilation when applied alone, far‐red light has impacts on underlying photochemistry which make it an especially relevant spectral region for photosynthesis. Although we did not measure enhancement at narrowband wavelengths greater than 720 nm, it is plausible that still greater enhancement could have been achieved at longer wavelengths as observed in spinach chloroplasts presented with 730 nm irradiance with a narrowband red background (Govindjee et al. 1964). Because those authors did not test beyond 730 nm where enhancement was still found to be greatest, the far‐red wavelengths around which enhancement ceases remains to be known. Note, however, that the narrowband light used in enhancement (and similar) studies is not normally monochromatic as if produced by a light source such as a laser, but will include a spread of wavelengths that are defined in terms of the peak wavelength. Enhancement effects obtained with such long wavelengths defined as 730 nm may well be due to the presence of other shorter and more strongly absorbed wavelengths within the nominal 730 nm spectrum.

In contrast to the findings of Govindjee et al. (1964), Zhen and Bugbee (2020) observed declining enhancement in lettuce as the wavelength of supplementary far‐red irradiance (added to a white light background) increased from 711 to 723 and 746 nm. There could be several explanations for these apparent differences, including growth spectrum and hence acclimation state (in greenhouse‐grown rice enhancement by far‐red light was greater when the growth spectrum included far‐red supplementation; Huber et al. 2024), the extent of absorption in the far‐red region, and the spectrum of actinic background light. With respect to the latter, the extent of enhancement observed in the present study may differ from that in actual sunlight or shade. Nonetheless, the results presented here are supportive of a role for far‐red light in ‘white light enhancement’ and consistent with a growing body of evidence on the subject (Zhen and van Iersel 2017; Zhen and Bugbee 2020; Huber et al. 2024). From a practical perspective, while the definition that PAR lies between 400 and 700 nm will suffice in some instances, it can lead to substantial error in work where actinic light spectrum contains far‐red light and in which absolute values or performance metrics are of interest, such as in modelling work, quantum yield studies, or the determination of horticultural light fixture performance (see Zhen et al. 2021). From an ecophysiological perspective, the enhancement in the SHADE spectrum may be significant because most leaves are exposed to shade beneath the canopy; these leaves may be surprisingly photosynthetically efficient in spectra which are classically regarded as poor for efficient photosynthesis.

We also show that enhancement correlates with the operational state of PSII and PSI, including Φ_PSII, qP, Fv'/Fm', Φ_PSI, and state transitions (measured as Fm'/Fm). These results show that the responses of photosynthesis at low irradiance are complex and understanding them depends on measuring not only assimilation but also the operation and regulation of PSI and PSII. Applying analyses of the kind we used here to leaves with strong irradiance gradients would need analyses of photosynthetic parameters that allow light penetration effects to be monitored, such as by using more excitation wavelengths for chlorophyll fluorescence. More generally our results agree closely with those of Hogewoning et al. (2012) and well with those of Mattila et al. (2020), implying that there is a general pattern of spectral responses of photosynthesis in C3 leaves. The biggest difference seems to lie in the 400–550 nm spectral region where light‐absorption by non‐photosynthetic pigments becomes important.

Supporting information

supplementary_material_05012025.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.