The LNK Gene Family: At the Crossroad between Light Signaling and the Circadian Clock

Light signaling pathways interact with the circadian clock to help organisms synchronize physiological and developmental processes to periodic environmental cycles. The plant photoreceptors responsible for clock resetting have been characterized, but signaling components that link the photoreceptors to the clock remain to be identified. Members of the family of NIGHT LIGHT–INDUCIBLE AND CLOCK-REGULATED (LNK) genes play key roles linking light regulation of gene expression to the control of daily and seasonal rhythms in Arabidopsis thaliana. Particularly, LNK1 and LNK2 were shown to control circadian rhythms, photomorphogenic responses, and photoperiod-dependent flowering time. Here we analyze the role of the four members of the LNK family in Arabidopsis in these processes. We found that depletion of the closely related LNK3 and LNK4 in a lnk1;lnk2 mutant background affects circadian rhythms, but not other clock-regulated processes such as flowering time and seedling photomorphogenesis. Nevertheless, plants defective in all LNK genes (lnkQ quadruple mutants) display developmental alterations that lead to increased rosette size, biomass, and enhanced phototropic responses. Our work indicates that members of the LNK family have both distinctive and partially overlapping functions, and are an essential link to orchestrate light-regulated developmental processes.


Introduction
The circadian clock is an endogenous timekeeper that enables plants to synchronize biological processes with daily and seasonal environmental conditions, allowing the allocation of resources during the most beneficial times of the day and the year. This anticipation of daily and seasonal environmental cycles allows plants to optimize internal processes in relation to external conditions, thus providing a fitness advantage [1][2][3][4][5][6].
Here we show that, unlike the RVE8 clade, the lnk1;lnk2;lnk3;lnk4 quadruple mutants (lnkQ) display a 1.3-h and 4-h period lengthening in comparison to lnk1;lnk2 double mutants and WT plants, respectively, while lnk3;lnk4 double mutants (lnk3;4) do not display circadian alterations at all. In addition, lnk quadruple mutants display altered leaf shape, bigger rosettes, and significantly enhanced biomass accumulation compared to the lnk1;lnk2 double mutants (lnk1;2), resembling what is seen with the rve multiple mutants [36]. We also found a strong hyponastic phenotype in lnk1;lnk2;lnk3;lnk4 quadruple mutants, which is not present in lnk1;lnk2 double mutants and is likely due to alterations in phototropic responses. Taken together, our findings provide another example of alterations in circadian clock components that result in enhanced plant growth and biomass accumulation, independently of their effect on flowering time, unveiling new paths for crop improvement.

Circadian Leaf Movement Analysis
For leaf movement analysis, plants were grown under 16-h light/8-h dark cycles until the appearance of the first pair of leaves. This period is referred to as the entrainment period. In order to measure circadian rhythms in leaf movement, plants were transferred to continuous white light (20 µmol·m −2 s −1 ) at 22 • C. The position of the first pair of leaves was recorded every 2 h for 5-6 days using digital cameras and the leaf angle was determined using ImageJ software [37]. Period estimates were calculated with Brass 3.0 software (Biological Rhythms Analysis Software System, available from http://www.amillar.org) and analyzed with Fast Fourier Transform Non-linear Least Squares (FFT-NLLS) using Brass 3.0 software. These experiments were performed in triplicate, with n = 8 for each genotype. The statistical analysis was done using a two-tailed Student's t-test.

Bioluminescence Assays
For bioluminescence assays, WT and LNK mutants were transformed with the pCCA1::LUC reporter using the floral dip method [38]. Seedlings were grown directly on half strength Murashige and Skoog 0.8% agar medium supplemented with 1% sucrose in a 96-well plate. One seed was placed per well and the seedlings were entrained under 16-h light/8-h dark cycles. After ≈7 days, 40 µL of luciferin (20 mM) were added to each well. The plate was transferred to constant light conditions and placed in a microplate luminometer LB-960 (Berthold Technologies, Bad Wildbad, Germany) to measure the bioluminescence emitted by each seedling every hour. After 5-6 days, data analysis was conducted using the Mikrowin 2000 software (version 4.29, Labsis Laborsysteme GmbH, Neunkirchen-Seelscheid, Germany). Period estimates were calculated with Brass 3.0 software and analyzed using FFT-NLLS. These experiments were performed in duplicate with n = 6 for each genotype. The statistical analysis was done using a two-tailed Student's t-test.

Flowering Time Analysis
For flowering time experiments, the plants were grown on soil at 22 • C under standard long days (LD; 16-h light/8-h dark cycles; 80 µmol·m −2 s −1 of white light), 12:12 days (12-h light/12-h dark cycles; 80 µmol·m −2 s −1 of white light), or short days (SD; 8-h light/16-h dark cycles; 140 µmol·m −2 s −1 of white light) depending on the experiment. Flowering time was estimated by counting the number of rosette leaves at the time of bolting. These experiments were performed in triplicate with n = 12 for each genotype. The statistical analysis was done using a two-tailed Student's t-test.

Hypocotyl Length Characterization
For hypocotyl length measurements, seedlings were grown on 0.8% agar under complete darkness, continuous red light (10 µmol·m −2 s −1 ), continuous blue light (2 µmol·m −2 s −1 ), continuous white light (LL), or under cycles of white light in SD or LD (all white light treatments 1 µmol·m −2 s −1 ). The final length of the hypocotyls was measured 4 days after germination. Light effects on hypocotyl elongation under continuous red and blue light were calculated normalizing hypocotyl length under each light regime to the hypocotyl length of the same genotype under constant dark conditions. For the white light experiments, absolute values of hypocotyl elongation are shown. These experiments were performed in quadruplicate with n = 20 for each genotype. The statistical analysis was done using a two-tailed Student's t-test.

Quantitative Real Time-Polymerase Chain Reaction
For time course analysis, 15-day old plants were grown in Murashige and Skoog 0.8% agar medium under 12-h light/12-h dark cycles at 22 • C, and then transferred for 3 days to continuous white light at 22 • C. Samples were collected every 4 h for 1 day, starting 24 h after the transfer to constant conditions. Total RNA was extracted using TRIzol reagent (Invitrogen, Carlsbad, CA, USA). One microgram of RNA was treated with RQ1 RNase-Free DNase (Promega, Madison, WI, USA) and subjected to retro-transcription with Super Script II Reverse Transcriptase (SSII RT) (Thermo Fisher Scientific, Waltham, MA, USA) and oligo-dT according to manufacturer's instructions. cDNAs were then amplified with FastStart Universal SYBR Green Master (Roche, Basel, Switzerland) using the Mx3000P Real Time PCR System (Agilent Technologies, Santa Clara, CA, USA) cycler. Samples were pooled across the time series before measuring mRNA abundance to estimate the difference in absolute mRNA levels among genotypes, rather than differences resulting from alterations in the timing of expression of each gene evaluated. The PP2A (AT1G69960) transcript was used as a housekeeping gene. Quantitative RT-PCR (qRT-PCR) analysis was conducted using the standard curve method as described in the Methods and Applications Guide from Agilent Technologies. Primer sequences and PCR conditions are available on request. Four biological samples were measured for each genotype.

Biomass and Leaf Morphology Analysis
To analyze leaf morphology, plants were grown on soil at 22 • C in LD conditions (16 h of light/8 h of dark; 80 µmol·m −2 s −1 white light). When plants were approximately 30-day old, the eighth leaf was cut, scanned, and analyzed with ImageJ [37], and the ratio blade/whole leaf length was calculated. These experiments were performed in triplicate, with n = 10 for each genotype. For size and biomass measurements, plants were grown on soil at 22 • C in SD conditions (8 h of light/16 h of dark; 140 µmol·m −2 s −1 white light). After bolting, rosettes were photographed and their perimeter determined digitally by drawing the smallest circle that covered the whole rosette using ImageJ [37]. Total aerial vegetative biomass (inflorescence stems removed) was quantified after drying the plants for 3 days at 40 • C in paper wraps. These experiments were performed in duplicate, with n = 18 for each genotype. All statistical analysis was done using a two-tailed Student's t-test.

Shade Avoidance Syndrome Assay
Plants were grown on soil at 22 • C in LD conditions (16 h of light/8 h of dark; 80 µmol·m −2 s −1 white light) with (treatment) or without (control) a 15-min pulse of far red light, given after the lights were turned off (end-of-day far-red treatment). When plants had 12 rosette leaves, the eighth leaf was cut, scanned, and measured using ImageJ [37]. For hypocotyl length measurements under simulated vegetative shade, seeds were sown on plates containing agar 0.8%, stratified and germinated under LD conditions (16 h of light/8 h of dark; 10 µmol·m −2 s −1 white light) for 2 days, then transferred to white light conditions with a neutral filter (control) or to white light with a green filter that generated a 0.35 red/far red ratio (treatment). The final length of the hypocotyls was measured after 4 days. Light effects on hypocotyl elongation are shown as absolute values. These experiments were performed in triplicate, with n = 10 for each genotype. The statistical analysis was done using a two-tailed Student's t-test.

Phototropic Response Characterization in Adult Plants and Seedlings
To measure the angle between rosette leaves, plants were grown in soil at 22 • C under SD conditions (8 h of light/16 h of dark; 140 µmol·m −2 s −1 white light). Six-week old plants were photographed from the side. The angle between the two opposite leaves closer to the light source was measured using ImageJ [37]. These experiments were performed in duplicate, with n = 18 for each genotype. For the phototropic reorientation of leaf blades, plants were grown on soil under LD conditions (16 h of light/8 h of dark; 80 µmol·m −2 s −1 white light). After the first pair of leaves was fully expanded, plants were exposed to a lateral source of white light (50 µmol·m −2 s −1 ). A leaf was considered to have reoriented if the leaf blade was perpendicular to the lateral light source. The number of reoriented leaves was counted in 4-week old plants. These experiments were performed in duplicate, with n = 12 for each genotype. For the characterization of phototropic responses in young seedlings, plants were germinated on vertical plates for 3 days in darkness, and then illuminated from one side with 1 µmol·m −2 s −1 of blue light for 8-10 h. The plates were scanned to measure the bending angle [39]. These experiments were performed in duplicate, with n = 4 for each genotype. All statistical analysis was done using a two-tailed Student's t-test.

Loss of Function of LNK3 and LNK4 Enhances the lnk1;2 Circadian Clock Phenotype
We previously described several clock-and light-regulated phenotypes for the lnk1-1;lnk2-1 double mutant. In order to study the degree of redundancy of the LNK family, we identified plants with T-DNA insertions in the LNK3 (SALK_085551C) and LNK4 (GK_846C06) loci, and we obtained the triple lnk1-1;lnk2-1,lnk3-1 (hereinafter referred to as lnk1;2;3) and lnk1-1;lnk2-1,lnk4-1 (hereinafter referred to as lnk1;2;4) mutants, as well as the quadruple lnk1-1;lnk2-1;lnk3-1;lnk4-1 (hereinafter referred to as lnkQ) mutant. To assess the functional redundancy among LNK genes on the function of the clock, we monitored circadian rhythms of leaf movement in WT plants, lnk3, lnk4, lnk3;4, lnk1;2, lnk1;2;3, and lnk1;2;4 mutants using time-lapse photography ( Figure 1A). Depletion of LNK3, LNK4, or both did not affect circadian rhythms. As shown previously, lnk1;2 mutants presented a longer circadian period than the WT (27.2 h vs. 24.5 h, respectively) [29]. Nevertheless, the lnk1;2 long period phenotype was not affected by depletion of neither LNK3 nor LNK4 genes. Interestingly despite showing clear leaf oscillations, an enhanced phototropic response exhibited by the lnkQ mutants generated a strong hyponastic phenotype that impaired the quantitative analysis of leaf movements in these plants. In order to surpass this issue, we transformed WT, lnk1;2, lnk3;4, and lnkQ mutants with a bioluminescent reporter and measured oscillations in gene expression driven by the CCA1 promoter. We found that the circadian period of CCA1 expression was lengthened in lnk1;2 (26.9 h) and lnkQ (28.4 h) mutants with respect to WT plants (25.4 h). Similar to what we observed for leaf movements, lnk3;4 double mutants did not display any alterations in the circadian expression of CCA1 (25.8 h) ( Figure 1B). These results indicate that LNK1 and LNK2 played a predominant role in the control of circadian rhythms compared to LNK3 and LNK4, which appeared to play a minor role that was only revealed in the absence of LNK1 and LNK2. the control of circadian rhythms compared to LNK3 and LNK4, which appeared to play a minor role that was only revealed in the absence of LNK1 and LNK2.

LNK3 and LNK4 Depletion in the lnk1;2 Mutant Background Does Not Affect Flowering Time
Another major physiological process that depends on the interaction between the circadian clock and light signaling is the day-length-dependent regulation of flowering time [40]. We already reported that lnk1;2 double mutants flowered later than WT plants under long-day photoperiods (16 h light/8 h dark). In order to assess the role, if any, of other LNK family members on flowering time regulation, we examined leaf number at bolting in lnk3, lnk4, lnk1;2, lnk3;4, lnk1;2;3, lnk1;2;4, and lnkQ mutants. lnk1;2 and lnkQ, but not lnk3;4 mutants, flowered later than WT plants under either 16:8 h or 12:12 h light:dark photoperiods, whereas no differences in flowering time were observed between lnk1;2 and lnkQ mutants (Figure 2A,B). On the other hand, no delay in flowering time was observed for the lnk3 and lnk4 single mutants compared to WT plants, or in lnk1;2;3 and lnk1;2,4 triple mutants compared to lnk1;2 double mutants under a 12:12 photoperiod (Supplementary Figure S1). Finally, none of the mutants analyzed displayed alterations in flowering time under short days (8 h light/16 :LUC activity measured for 5 days in LL, after entrainment under long-day conditions. Bioluminescence was recorded every 2 h over 6 days. Periods of circadian rhythms in leaf movement and CCA1::LUC activity were estimated with BRASS 3.0 software (Biological Rhythms Analysis Software System). Error bars indicate standard error of the mean (SEM). Student's t-test was performed between mutants and wild-type (WT) (*: significantly different, p ≤ 0.05; **: significantly different, p ≤ 0.01; ***: significantly different, p ≤ 0.001; ns: not significant).

LNK3 and LNK4 Depletion in the lnk1;2 Mutant Background Does Not Affect Flowering Time
Another major physiological process that depends on the interaction between the circadian clock and light signaling is the day-length-dependent regulation of flowering time [40]. We already reported that lnk1;2 double mutants flowered later than WT plants under long-day photoperiods (16 h light/8 h dark). In order to assess the role, if any, of other LNK family members on flowering time regulation, we examined leaf number at bolting in lnk3, lnk4, lnk1;2, lnk3;4, lnk1;2;3, lnk1;2;4, and lnkQ mutants. lnk1;2 and lnkQ, but not lnk3;4 mutants, flowered later than WT plants under either 16:8 h or 12:12 h light:dark photoperiods, whereas no differences in flowering time were observed between lnk1;2 and lnkQ mutants (Figure 2A,B). On the other hand, no delay in flowering time was observed for the lnk3 and lnk4 single mutants compared to WT plants, or in lnk1;2;3 and lnk1;2,4 triple mutants compared to lnk1;2 double mutants under a 12:12 photoperiod (Supplementary Figure S1). Finally, none of the mutants analyzed displayed alterations in flowering time under short days (8 h light/16 h dark). These findings suggest that LNK1 and LNK2 were specifically required for the day length-dependent flowering time pathway, rather than for the control of the transition from vegetative to reproductive growth per se, whereas LNK3 and LNK4 play no role in the control of the floral transition.
h dark). These findings suggest that LNK1 and LNK2 were specifically required for the day lengthdependent flowering time pathway, rather than for the control of the transition from vegetative to reproductive growth per se, whereas LNK3 and LNK4 play no role in the control of the floral transition.

The LNK Family Controls Seedling Photomorphogenesis
Besides flowering time, hypocotyl elongation, an early developmental phenomenon, was also regulated by light and the circadian clock. We first analyzed hypocotyl elongation in darkness. lnk1;2 and lnkQ mutants were only slightly shorter than WT seedlings, while the lnk3;4 double mutants

The LNK Family Controls Seedling Photomorphogenesis
Besides flowering time, hypocotyl elongation, an early developmental phenomenon, was also regulated by light and the circadian clock. We first analyzed hypocotyl elongation in darkness. lnk1;2 and lnkQ mutants were only slightly shorter than WT seedlings, while the lnk3;4 double mutants showed no differences (Supplementary Figure S2), suggesting a minor role for LNK1 and LNK2 in the regulation of cell elongation. We then studied the inhibition of hypocotyl elongation under different light treatments. The lnk1;2 and lnkQ mutants displayed longer hypocotyls than WT seedlings under continuous blue and red light, indicating hyposensitivity to both wavelengths. Nevertheless, no differences were observed between the double and quadruple mutants, nor between the lnk3;4 double mutants and WT plants ( Figure 3A,B). We then assessed the effect of different photoperiods on hypocotyl elongation. Interestingly, both lnk1;2 and lnkQ mutants exhibited longer hypocotyls than WT or lnk3;4 seedlings under all the photoperiodic conditions evaluated. No significant differences were detected between lnk1;2 and lnkQ, nor between lnk3;4 and WT seedlings ( Figure 3C-E). Finally, we evaluated light inhibition of hypocotyl elongation under continuous white light. Under this condition, all lnk mutants displayed significantly longer hypocotyls than WT plants, with lnk1;2 and lnkQ showing longer hypocotyls than lnk3;4 mutants ( Figure 3F). Collectively, these results indicate that LNK1 and LNK2 have a predominant role in the mediation of light inhibition of hypocotyl growth, while LNK3 and LNK4 appear to have a minor, if any, role in this process. This idea is further supported by the fact that lnk1;2;3 and lnk1;2;4 triple mutants displayed similar inhibition of hypocotyl growth compared to lnk1;2 and lnkQ under several photoperiods, and by the observation that lnk3 and lnk4 simple mutants did not exhibit any alteration at all in hypocotyl elongation (Supplementary Figure S3).

Cumulative Effects of LNK Depletion on the Expression of Core Circadian Clock Genes
Although depletion of LNK3 and LNK4 lengthened circadian period in the lnk1;2 mutant background, other light-and clock-regulated processes, such as flowering time and hypocotyl elongation, were not further affected in lnkQ compared to lnk1;2 mutants. This finding revealed a significant degree of redundancy for the LNK family in the function of the circadian clock, but a seemingly prevailing role for LNK1 and LNK2 in the control of light signaling pathways. To obtain a more detailed understanding of the role of the LNK family in the control of the circadian clock, we first analyzed the expression patterns of LNK3 and LNK4 genes under free running conditions. Similarly to what was reported for LNK1 and LNK2, LNK3 and LNK4 were also rhythmically expressed, with mRNAs showing peak levels in the subjective morning (Supplementary Figure S4) [29]. It is known that LNK1 and LNK2 promote the expression of PRR5 and TOC1, and ultimately these PRRs contribute to the downregulation of CCA1 and LHY at midnight [29,30]. We therefore measured the expression levels of four of the core clock genes in WT, lnk1;2, lnk3;4, and lnkQ mutants. Expression levels of TOC1 and PRR5 were reduced in lnk1;2 and lnkQ mutants compared to WT plants ( Figure 4A,B). The expression of the morning genes CCA1 and LHY was also reduced in both lnk1;2 and lnkQ mutants, but unaffected in lnk3;4 mutants ( Figure 4C,D). Interestingly, PRR5 and CCA1 expression was more severely inhibited in lnkQ than in lnk1;2 mutants, thus revealing some degree of redundancy of the LNKs in the control of gene expression for a sub-set of core clock components. Although depletion of LNK3 and LNK4 lengthened circadian period in the lnk1;2 mutant background, other light-and clock-regulated processes, such as flowering time and hypocotyl elongation, were not further affected in lnkQ compared to lnk1;2 mutants. This finding revealed a significant degree of redundancy for the LNK family in the function of the circadian clock, but a seemingly prevailing role for LNK1 and LNK2 in the control of light signaling pathways. To obtain a more detailed understanding of the role of the LNK family in the control of the circadian clock, we first analyzed the expression patterns of LNK3 and LNK4 genes under free running conditions. Similarly to what was reported for LNK1 and LNK2, LNK3 and LNK4 were also rhythmically expressed, with mRNAs showing peak levels in the subjective morning (Supplementary Figure S4) [29]. It is known that LNK1 and LNK2 promote the expression of PRR5 and TOC1, and ultimately these PRRs contribute to the downregulation of CCA1 and LHY at midnight [29,30]. We therefore measured the expression levels of four of the core clock genes in WT, lnk1;2, lnk3;4, and lnkQ mutants. Expression levels of TOC1 and PRR5 were reduced in lnk1;2 and lnkQ mutants compared to WT plants ( Figure 4A,B). The expression of the morning genes CCA1 and LHY was also reduced in both lnk1;2 and lnkQ mutants, but unaffected in lnk3;4 mutants ( Figure 4C, D). Interestingly, PRR5 and CCA1 expression was more severely inhibited in lnkQ than in lnk1;2 mutants, thus revealing some degree of redundancy of the LNKs in the control of gene expression for a sub-set of core clock components.  Mean expression levels were measured using quantitative PCR (qPCR) for the evening genes (a) TOC1 and (b) PRR5, as well as for the morning genes (c) CCA1 and (d) LHY. Four biological replicates were used for measuring mRNA abundance. The analysis was conducted after pooling samples collected every 4-h across a time-course series, in order to evaluate difference in absolute mRNA levels among genotypes, rather than differences resulting from alterations in the timing of gene expression. Error bars indicate SEM. Student's t-test was performed between mutants and WT (*: significantly different, p ≤ 0.05; **: significantly different, p ≤ 0.01; ***: significantly different, p ≤ 0.001; ns: not significant).

The LNK Family Acts as a Growth Modulator in the Vegetative Stage
In contrast to the apparent similarity for both flowering time and hypocotyl elongation between lnk1;2 and lnkQ mutants, these plants displayed significant differences in their adult phenotypes ( Figure 5A). To further investigate these phenotypic differences, we measured rosette perimeter in plants grown under short-day conditions. The rosettes of lnkQ mutants were larger than those of WT plants (31.5 cm vs. 21.3 cm respectively) and lnk1;2 mutants (31.5 cm vs. 25.3 cm respectively); those of lnk3;4 double mutants, on the other hand, were similar to WT plants (22.5 cm vs. 21.3 cm respectively) ( Figure 5B). We next evaluated biomass accumulation, measured as the dry weight of adult plants grown under short-day conditions. We found that the larger rosettes of lnkQ plants were associated with a 72% increase in dry weight compared to WT plants and a 27% increase compared to lnk1;2 mutants; the lnk1;2 double mutants exhibited a 36% increase compared to WT plants, whereas lnk3;4 showed no changes ( Figure 5C). Another trait affected in adult lnkQ mutants was leaf morphology ( Figure 5D). To quantify this, we calculated the leaf blade/total leaf length ratio and measured petiole length of the eighth leaf in plants grown under long-day conditions. All mutants assessed displayed a smaller ratio compared to that of WT plants, indicating smaller leaf lamina and/or longer petioles ( Figure 5E). These differences could be attributed to either a difference in leaf blade length, petiole length, or both. Petiole length was similar in lnkQ and lnk1;2 mutants (2.7 cm on average in both mutants) and longer than those of WT (1.7 cm), indicating that the smaller leaf blade/total leaf length ratio observed in lnkQ under long-day conditions was due to shorter leaf blades. On the other hand, lnk3;4 double mutants showed no differences in petiole length compared to WT ( Figure 5F). Taken together, these results indicate that the LNK co-activator family redundantly controls petiole length and leaf growth, which in turn impacts rosette size and biomass accumulation in adult plants. bars indicate SEM. Student's t-test was performed between mutants and WT (*: significantly different, p ≤ 0.05; **: significantly different, p ≤ 0.01; ***: significantly different, p ≤ 0.001; ns: not significant).

The LNK Family Acts as a Growth Modulator in the Vegetative Stage
In contrast to the apparent similarity for both flowering time and hypocotyl elongation between lnk1;2 and lnkQ mutants, these plants displayed significant differences in their adult phenotypes ( Figure 5A). To further investigate these phenotypic differences, we measured rosette perimeter in plants grown under short-day conditions. The rosettes of lnkQ mutants were larger than those of WT plants (31.5 cm vs. 21.3 cm respectively) and lnk1;2 mutants (31.5 cm vs. 25.3 cm respectively); those of lnk3;4 double mutants, on the other hand, were similar to WT plants (22.5 cm vs. 21.3 cm respectively) ( Figure 5B). We next evaluated biomass accumulation, measured as the dry weight of adult plants grown under short-day conditions. We found that the larger rosettes of lnkQ plants were associated with a 72% increase in dry weight compared to WT plants and a 27% increase compared to lnk1;2 mutants; the lnk1;2 double mutants exhibited a 36% increase compared to WT plants, whereas lnk3;4 showed no changes ( Figure 5C). Another trait affected in adult lnkQ mutants was leaf morphology ( Figure 5D). To quantify this, we calculated the leaf blade/total leaf length ratio and measured petiole length of the eighth leaf in plants grown under long-day conditions. All mutants assessed displayed a smaller ratio compared to that of WT plants, indicating smaller leaf lamina and/or longer petioles ( Figure 5E). These differences could be attributed to either a difference in leaf blade length, petiole length, or both. Petiole length was similar in lnkQ and lnk1;2 mutants (2.7 cm on average in both mutants) and longer than those of WT (1.7 cm), indicating that the smaller leaf blade/total leaf length ratio observed in lnkQ under long-day conditions was due to shorter leaf blades. On the other hand, lnk3;4 double mutants showed no differences in petiole length compared to WT ( Figure 5F). Taken together, these results indicate that the LNK co-activator family redundantly controls petiole length and leaf growth, which in turn impacts rosette size and biomass accumulation in adult plants.

lnkQ Mutants Show Normal Shade Avoidance but Enhanced Phototropic Responses at the Adult Stage
The long petioles and small leaf blades of fully grown lnk1;2 and lnkQ mutants resemble those of plants displaying the shade avoidance syndrome [41]. To assess this trait, we exposed WT and mutant plants to an end-of-day far red (FR) treatment, which simulated shade conditions. All genotypes were responsive to simulated shade, showing a smaller leaf blade/whole leaf length ratio than plants grown under control light conditions ( Figure 6A). Furthermore, the magnitude of the response was the same in all genotypes (Supplementary Figure S5A). In order to exclude effects of the life stage and the nature of the treatment used for shade simulation, we evaluated the effect of a continuous simulated shade condition under a long-day photoperiod on hypocotyl elongation. Once again, all the genotypes evaluated displayed identical responses to shade simulation, with longer hypocotyls under shade conditions compared to those of control seedlings ( Figure 6B and Supplementary Figure S5B). These data suggest that the phenotype of lnk1;2 and lnkQ mutants was not due to constitutive or enhanced responses to shade. Indeed, these findings are consistent with the fact that both mutants, grown under short-day photoperiods, displayed increased biomass accumulation, opposite to the biomass reduction triggered in response to shade signals [41].
Phototropism, perhaps the most recognizable adaptive growth response in plants, is the result of differential cell elongation, which results in the orientation of organs towards the light source to optimize photosynthetic light capture. In flowering plants, phototropism is induced via UV and blue wavelengths (290-500 nm), and modulated to some extent by red light. Remarkably, this phototropic response is evoked over a wide range of light intensities, ranging from extremely small amounts of light to the blue light intensity present on a sunny day under field conditions [42]. Whereas WT plants and lnk3;4 mutants presented flat rosettes, with angles between opposite leaves of 128 and 131 degrees respectively, lnk1;2 mutants displayed more erect leaves with an angle of 117 degrees. lnkQ leaves formed a notoriously acute angle of 80 degrees, indicating a strong hyponastic phenotype in this high order lnk mutant ( Figure 6C). Considering the fact that the mutants did not display constitutive or enhanced shade avoidance responses, we hypothesized that the phenotype could be due to an altered phototropic response. We thus scored the reorientation of rosette leaves in plants grown under vertical illumination until the first pair of leaves appeared, and then exposed only to a lateral light source. When plants had 8 to 10 leaves, we measured the proportion of leaves whose blades had re-oriented towards the light source. In WT plants, about half of the leaf blades (48%) adopted such a position, particularly those closer to the light source. Interestingly, all lnk mutants tested displayed a significant increase in the proportion of reoriented leaf blades: lnk3;4 and lnk1;2 mutants reoriented 60% and 72% of their leaves, respectively, but 87% of the lnkQ leaves adopted this arrangement ( Figure 6D). To assess phototropic responses at other developmental stages, we measured hypocotyl bending in response to a lateral blue light source, as proposed in Fankhauser et al. [39]. No differences were detected between WT and all lnk mutants using this experimental setup, suggesting that these genes redundantly modulate phototropic responses mainly in the adult stage (Supplementary Figure S6).

Discussion
Our understanding of the plant circadian clock has significantly changed in the last decade, from a relatively simple gene network running on mostly negative transcriptional feedback loops, to a model with several interlocked regulatory gears, in which both activators and repressors play major roles. The recent characterization of the RVE and LNK gene families has uncovered a new layer of clock activators. Here we analyzed the whole LNK family in Arabidopsis, and show that different members have distinctive, but also partially redundant, roles in the regulation of circadian rhythms, growth, and development. Significantly, the depletion of the entire family results in a noticeable circadian-period lengthening and remarkable photomorphogenic phenotypes, thus revealing an important role for this gene family as transducers of light signals to the clock.
One of the most important gene families controlling clock function is that of the single MYB transcription factors CCA1, LHY, and the RVEs, grouped in the REVEILLE/LHY-CCA1-LIKE (RVE/LCL) subfamily, are characterized by the presence of a unique LCL protein domain. (d) The proportion of leaf blades perpendicular to the light source was determined in plants grown in long days with vertical illumination until the first pair of leaves appeared, and then exposed to a lateral light source until the plants developed 8-10 leaves. Error bars indicate SEM. Student's t-test was performed between mutants and WT (*: significantly different, p ≤ 0.05; **: significantly different, p ≤ 0.01; ***: significantly different, p ≤ 0.001; ns: not significant).

Discussion
Our understanding of the plant circadian clock has significantly changed in the last decade, from a relatively simple gene network running on mostly negative transcriptional feedback loops, to a model with several interlocked regulatory gears, in which both activators and repressors play major roles. The recent characterization of the RVE and LNK gene families has uncovered a new layer of clock activators. Here we analyzed the whole LNK family in Arabidopsis, and show that different members have distinctive, but also partially redundant, roles in the regulation of circadian rhythms, growth, and development. Significantly, the depletion of the entire family results in a noticeable circadian-period lengthening and remarkable photomorphogenic phenotypes, thus revealing an important role for this gene family as transducers of light signals to the clock.
One of the most important gene families controlling clock function is that of the single MYB transcription factors CCA1, LHY, and the RVEs, grouped in the REVEILLE/LHY-CCA1-LIKE (RVE/LCL) subfamily, are characterized by the presence of a unique LCL protein domain. Interestingly, almost half of the members of this family belong to the RVE clade. While RVE8 and its homologs RVE4 and RVE6 function in a partially redundant manner to modulate the pace of the clock, RVE3 and RVE5 exhibit minor roles in circadian rhythm regulation. Nevertheless, rve3;rve4;rve5;rve6;rve8 quintuple mutant plants display significant growth alterations compared to WT plants and rve4;rve6;rve8 triple mutants. Here we explored the smaller family of LNK genes. Flowering plants have three main types of LNK genes: LNK1, LNK2, and LNK3/LNK4 clades. Homologs of LNK3/LNK4 are found as single copy genes in several plant species, but as duplicates in all crown Brassicaceae. They are likely the product of the most recent genome duplication at the base of this family, which may account for their redundant functions [29].
In a characterization of LNK1 and LNK2 as transcriptional co-activators of the circadian core oscillator, Xie et al. reported the interaction of both LNKs with CCA1, LHY, RVE4, and RVE8. LNK1 and LNK2 induce PRR5 transcription through their interaction with RVE4 and RVE8, which, unlike LNKs, have DNA binding domains [30]. Although the LNK family members lack known functional domains, they share two highly conserved regions called R1 and R2. Mutations in those motifs impair the ability of LNK1 to interact with RVE4 [30]. In addition, it was recently reported that all LNKs interact with MYB3, an R2R3-MYB transcription factor that inhibits the Arabidopsis phenylpropanoid biosynthesis through the repression of the cinnamate 4-hydroxylase (C4H) gene, a key enzyme involved in lignin and flavonoid biosynthesis [43]. Indeed, it is the C-terminal domain common to all LNKs, where the R1 and R2 sites are found, that mediates the interaction with MYB3, and the simultaneous mutation of both R1 and R2 in LNK1 impaired the repressor activity of MYB3 on a C4H reporter [43]. Although LNK3 and LNK4 proved unable to act as co-repressors of MYB3 [43], they have recently been shown to interact with RVE8 [31], and both LNK1 and LNK3 interact with the LCL domain of RVE8 [44]. All evidence considered, it is possible that in the absence of LNK1 and LNK2, RVE4 and/or RVE8 may interact with LNK3 and/or LNK4 to promote the expression of PRR5 and TOC1, which would explain the further period lengthening in the lnkQ background.
In contrast to what we have seen regarding the control of circadian rhythms, other clock-regulated biological processes, such as flowering time and the inhibition of hypocotyl elongation, were equally affected in both lnk1;2 and lnkQ mutants, suggesting that both LNK1 and LNK2 have a function that LNK3 and LNK4 lack. LNK3 and LNK4 are smaller than LNK1 and LNK2 (≈30 kDa and ≈60 kDa, respectively). Zhou et al. showed that an extra N-terminal tail (ENT) of ≈300 amino acids, present only in LNK1 and LNK2, is necessary for the co-repression of C4H by LNK1 and MYB3 [43]. In fact, when the ENT domain of LNK1 was fused to the N-terminus of LNK3, the ENT-LNK3 chimeric protein behaved like LNK1. The ENT domain may also be responsible for the repressive role of LNK1 on the expression of anthocyanin structural genes, such as UGT79B1, counteracting in this way the activating function of RVE8 [31]. LNK1 and LNK2 are therefore more complex proteins, whereas LNK3 and LNK4 may be more circumscribed to protein-protein interactions through the R1 and R2 regions common to all members of the family. These differences could explain the predominant roles of LNK1 and LNK2, compared to LNK3 and LNK4, in the regulation of flowering time and the inhibition of hypocotyl elongation.
It is interesting to notice that, in contrast to the case of rve4;rve6;rve8 and rve3;rve4;rve5;rve6;rve8 mutants [36], almost every morphologic trait analyzed here was enhanced in the lnkQ mutants compared to lnk1;2. Both the deletion of RVEs or LNKs trigger an increase in biomass accumulation in comparison to WT plants; these mutant combinations display, however, completely different phenotypes. It was recently shown that LNKs, but not RVE8, interact with the RNA polymerase II and the transcript elongation FACT (facilitates chromatin transactions) complex to rhythmically co-occupy PRR5 and TOC1 loci and regulate their circadian transcription. This interaction seems to be important to modulate the transcript initiation and elongation, and is mediated by the R1 and R2 regions [44]. It is thus plausible that the proper transcription of other genes besides PRR5 and TOC1 is regulated by the whole LNK family, accounting for such a pleiotropic phenotype.
We found that adult lnkQ quadruple mutants have enhanced phototropic responses. Previous work showed that the expression of both LNK1 and LNK2 is induced by active phytochromes [29]. In etiolated seedlings, phytochromes do not detect the light gradient per se but they do fine-tune the magnitude of the phototropic response. Phytochromes, with a predominant function of phytochrome A, enhance phototropism modulating the phototropin signaling pathway at several steps in etiolated seedlings [45][46][47]. More recently, it has been shown that shade-induced low red to far-red (R/FR) ratio perceived by phytochrome B enhances the phototrophic response in green de-etiolated seedlings through the action of PIF genes and the auxin biosynthetic YUCCA genes [48]. The work by Goyal et al. demonstrates that different light signaling pathways modulate phototropic responses in different developmental stages, highlighting the complex crosstalk between the R/FR sensing phytochromes and the blue-light-sensing phototropins [48]. It is therefore possible that LNKs mediate signaling downstream of phytochromes, acting as co-activators and/or co-repressors of the phototropic responses mediated by PIFs. In fact, it was recently shown that the ELONGATED HYPOCOTYL 5 (HY5) transcription factor, a signaling hub downstream of several photoreceptors and a key mediator of photomorphogenesis, binds specifically to the promoters of all LNK genes [49].
New insights into circadian clock regulation are emerging as promising avenues for crop improvement. It has been recently shown that an almost complete deletion of the LNK2 homolog is responsible for the intriguing delay in circadian period observed in domesticated tomato [50,51]. It has been hypothesized that a longer clock period may have been selected as this crop expanded from its center of origin near the equator to higher latitudes. However, the clock pace in local tomato landraces around the equator is also modified compared to sympatric wild ancestors. Considering that biomass increase is a remarkable feature of lnk mutants in Arabidopsis, it would be interesting to explore whether the circadian phenotype of cultivated tomatoes may be an unintended result of selection on biomass or related traits, that later facilitated the adjustment to different photoperiods. Whereas the increased biomass in plants exhibiting alterations in the circadian clock function may be the gateway to new agronomic applications, it remains to be seen whether this growth promotion brings collateral consequences in other aspects, as proposed in the classical dilemma to grow or defend [52].
Supplementary Materials: The following are available online at http://www.mdpi.com/2073-4425/10/1/2/s1, Figure S1: Characterization of LNK3 and LNK4 in the control of flowering, Figure S2: Role of the LNK family in the control of hypocotyl elongation, Figure S3: Characterization of LNK3 and LNK4 in the control of photomorphogenesis, Figure S4: Circadian expression of LNK3 and LNK4 genes, Figure S5: LNK mutants respond normally under shade conditions, Figure S6: lnk mutant seedlings exhibit normal phototropic responses.