Structure-activity relationship of 2,4-D correlates auxin activity with the induction of somatic embryogenesis in Arabidopsis thaliana

2,4-dichlorophenoxyacetic acid (2,4-D) is a synthetic analogue of the plant hormone auxin that is commonly used in many in vitro plant regeneration systems, such as somatic embryogenesis (SE). Its effectiveness in inducing SE, compared to the natural auxin indole-3-acetic acid (IAA), has been attributed to the stress triggered by this compound rather than its auxin activity. However, this hypothesis has never been thoroughly tested. Here we used a library of 40 2,4-D analogues to test the structure-activity relationship with respect to the capacity to induce SE and auxin activity in Arabidopsis thaliana. Four analogues induced SE as effectively as 2,4-D and 13 analogues induced SE but were less effective. Based on root growth inhibition and auxin response reporter expression, the 2,4-D analogues were classified into different groups, ranging from very active auxins to not active. A halogen at the 4-position of the aromatic ring was important for auxin activity, whereas a halogen at the 3-position resulted in reduced activity. Moreover, a small substitution at the carboxylate chain was tolerated, as was extending the carboxylate chain with two but not with one carbon. In the process, we also identified two 2,4-D analogues as efficient inducers of adventitious root formation and several possible anti-auxins. The auxin activity of the 2,4-D analogues was consistent with their simulated TIR1-Aux/IAA coreceptor binding characteristics. A strong correlation was observed between SE induction efficiency and auxin activity, indicating that the stress-related effects triggered by 2,4-D that are considered important for SE induction are down-stream of auxin signaling.

Omid Karami 1* , Hanna de Jong 2* , Victor J. Somovilla 3 , Beatriz Villanueva Acosta 1 , Aldo Bryan 1 6 Abstract 1 7 2,4-dichlorophenoxyacetic acid (2,4-D) is a synthetic analogue of the plant hormone auxin that is 1 8 commonly used in many in vitro plant regeneration systems, such as somatic embryogenesis (SE). Its 1 9 effectiveness in inducing SE, compared to the natural auxin indole-3-acetic acid (IAA), has been 2 0 attributed to the stress triggered by this compound rather than its auxin activity. However, this 2 1 hypothesis has never been thoroughly tested. Here we used a library of 40 2,4-D analogues to test the 2 2 structure-activity relationship with respect to the capacity to induce SE and auxin activity in 2 3 Arabidopsis thaliana. Four analogues induced SE as effectively as 2,4-D and 13 analogues induced 2 4 SE but were less effective. Based on root growth inhibition and auxin response reporter expression, 2 5 the 2,4-D analogues were classified into different groups, ranging from very active auxins to not 2 6 active. A halogen at the 4-position of the aromatic ring was important for auxin activity, whereas a 2 7 halogen at the 3-position resulted in reduced activity. Moreover, a small substitution at the 2 8 carboxylate chain was tolerated, as was extending the carboxylate chain with two but not with one 2 9 carbon. In the process, we also identified two 2,4-D analogues as efficient inducers of adventitious 3 0 root formation and several possible anti-auxins. The auxin activity of the 2,4-D analogues was 3 1 consistent with their simulated TIR1-Aux/IAA coreceptor binding characteristics. A strong correlation 3 2 was observed between SE induction efficiency and auxin activity, indicating that the stress-related 3 3 effects triggered by 2,4-D that are considered important for SE induction are down-stream of auxin 3 4 signaling. 3 5 The plant hormone auxin plays a central role in the development of plants. In the 1930s, the structure 4 1 of the natural auxin indole-3-acetic acid (IAA) was first described (Ma et al., 2018). A few years later, 4 2 during WWII, 2,4-dichlorophenoxyacetic acid (2,4-D) was discovered as a synthetic auxin analogue 4 3 that can be used as a herbicide, targeting dicots. Today, it is still broadly used as such in gardens and 4 4 in agriculture (Peterson et al., 2016). Apart from the cell elongation promoting effect, which led to its 4 5 discovery as auxin analogue, 2,4-D acts differently in various physiological and molecular assays Mitao and Kakimoto, 2014). 4 8 Binding of IAA to its receptors triggers a transcriptional response. The classical nuclear IAA 4 1 9 7 Specific 2,4-D analogues as tools to modulate root system architecture 1 9 8 When establishing the root growth inhibition of our 2,4-D analogues, we observed that several 1 9 9 compounds had a unique effect on the root system architecture (RSA). The RSA of a plant describes 2 0 0 the organization of the primary, lateral and adventitious roots. This includes root hairs that increase 2 0 1 the surface area and thus promote the uptake of water and nutrients (Smith and de Smet, 2012 All 2,4-D analogues that were designated as very active, active and weakly active positively increased 2 0 8 the number of root hairs in Arabidopsis seedling roots (not shown). Specifically, we observed a strong 2 0 9 effect on root hair formation on root tissues treated with 5 µM of 3-Cl, 3,4-D, 2,4-DP or Mecoprop.
We found that the root hair formation was dose-dependent for all four analogues (Fig. S4).
Interestingly, treatment with 0,5 or 1 µM of the weakly active 3-Cl did not lead to a strong inhibition 2 1 2 of root growth, like with the other three active or very active compounds (Fig. S4), whereas it still 2 1 3 strongly enhanced root hair development. This result reflects the weaker auxin response induced by 3-2 1 4 Cl in the root meristem, leading to reduced root growth inhibition (Fig. 2), while it still induces a 2 1 5 relatively strong auxin response in the differentiation zone, resulting in ectopic root hair formation 2 1 6 ( Fig. 4).
We also observed that 3-Cl and 2,5-D significantly promoted the number of lateral roots, whereas 2,4-2 1 8 D and other analogues with strong auxin activity initially induced many lateral root meristems, but 2 1 9 these meristems quickly deteriorated into amorphous callus (Fig. 5A). As lateral root and adventitious 2 2 0 root (AR) induction are highly linked, we tested the capacity of different concentrations of 2,4-D, 2 2 1 IBA, 3-Cl and 2,5-D in AR induction from hypocotyls of dark grown Arabidopsis seedlings. AR 2 2 2 induction is a crucial process in clonal crop propagation by cuttings or shoot regeneration, and is well- known to be promoted in many plant species by the natural auxin indole-3-butyric acid (IBA). As 2 2 4 with lateral roots and in line with previous observations (da Costa et al., 2018), treatment with 2,4-D 2 2 5 produced a low number of ARs at low µM concentrations and only undesired callus at higher µM 2 2 6 concentrations ( Fig. 5B,C). Treatment with 3-Cl and 2,5-D, however, efficiently induced ARs at 5 µM 2 2 7 (Fig. 5B, C). In addition, the number of ARs induced by 3-Cl and 2,5-D was significantly higher 2 2 8 compared to IBA treatment at a similar concentration and also compared to 57 µM IAA or 2 µM of 2 2 9 the synthetic auxin 1-naphthaleneacetic acid (NAA) (Fig. 5D), treatments that have previously been 2 3 0 shown to efficiently induces ARs from of Arabidopsis hypocotyls (da Costa et al., 2018). With these 2 3 1 results, we can conclude that 3-Cl and 2,5-D are excellent candidates for inducing AR formation in 2 3 2 Arabidopsis, and that despite their structural similarity with 2,4-D, they show a unique biological 2 3 3 activity. This probably relates to their mild and specific activity as auxin analogue, as reflected by the 2 3 4 specific expression pattern of the pDR5:GUS reporter following treatment with these compounds. the 2,4-D analogues for which we observed no clear effect on root length or pDR5 activity, namely 3-2 3 7 Me, 2,3-D, 3,5-D, 2,4,6-T, and 2,4-DnP, did significantly inhibit lateral root formation, with some 2 3 8 having a stronger effect (Fig. 6A, B). In order to investigate the lateral root inhibition-responsiveness 2 3 9 to these potential anti-auxins, we examined the effect of short term treatment (1 day) on the different 2 4 0 developmental stages of lateral root formation, using the pDR5:GUS reporter activity as marker shown). This is in line with the observation that treatment with these compounds did not lead to a 2 4 5 reduced expression of the pDR5:GUS reporter in the main root tip (Fig. S2). However, these 2,4-D 2 4 6 derived anti-auxin candidates did reduce the number of stage 1 primordia, had no effect on the 2 4 7 number of stage 3 primordia, whereas they had a differential effect on the number of stage 2 2 4 8 primordia and stage 4 lateral roots (Fig. 6C). The effect of these compounds on lateral root formation 2 4 9 might indeed be caused by their activity as anti-auxins, however we cannot exclude that they affect 2 5 0 other processes, such as auxin transport. 2 5 1 Based on the combined results we now classified all tested 2,4-D analogues from our library into five 2 5 2 groups, namely very active, active, weakly active auxin analogues and not active and anti-auxin 2 5 3 compounds (Fig. S2). Our results show that the capacity of these auxin analogues to induce SE is 2 5 4 tightly linked to their auxin activity.  (Table 1).
Among the library entries tested in silico are alpha substituted compounds (methyl, ethyl or the 2 6 3 Mecoprop derivatives 2,4-DP, 2,4-DB and Mecoprop, respectively) that due to their stereocenter can 2 6 4 occur as the enantiomeric R or S stereoisomers. These compounds were tested as mixture of these 2 6 5 enantiomers in the in planta assays. The molecular dynamics simulation uniquely allowed us to 2 6 6 investigate at a molecular level the potential differential binding of each enantiomer (notated as their 2 6 7 acronym with an added R and S).
In the evaluation of the binding capacity of each auxin analogue within the system, we studied 2 6 9 different aspects: the enthalpy of the system, the root-mean-square deviation (RMSD) of atomic 2 7 0 positions of TIR1 along the simulation trajectory, the hydrogen bond network established by the auxin 2 7 1 analogue and the distance between the auxin analogue and the proline (7) of the Aux/IAA peptide 2 7 2 ( Table 1). The RMSD of TIR1 reflects the system stability during the simulation. The last parameter, 2 7 3 the distance (gamma) between the auxins analogue and the proline (7) of the Aux/IAA peptide, would 2 7 4 show us which analogue is able to establish a CH·π interaction with the degron peptide that 2 7 5 contributes to the binding interactions and thus to the stability of the complex (Wang and Yao, 2019).
Our molecular dynamics simulations were based on a previously published method (Hao and Yang, 2 7 7 2010), but with a much longer trajectory of 200 ns instead of 2.5 ns.
The enthalpy was calculated using MMPBSA (Molecular Mechanics Poisson-Boltzmann Surface 2 7 9 Area) with a lower enthalpy value indicating better binding. All the compounds classified as very 2 8 0 active auxin analogues showed an enthalpy value below -12 Kcal/mol. Surprisingly, the weakly active 2 8 1 2,6-D and inactive 4-NO 2 were also in this range, suggesting that the binding enthalpy is not the sole 2 8 2 parameter predictive for auxin activity. The TIR1 protein RMSD calculation was carried out over the 2 8 3 amino acids from position 50 till the end of the protein, because the first 50 amino acids showed extra 2 8 4 flexibility. The TIR1 N-terminal part is normally stabilized by the interaction with the SKP protein, 2 8 5 which we did not include in our analysis. Moreover this part does not interact with the auxin analogue 2 8 6 induce SE from Arabidopsis IZEs, which indicates that the capacity to induce SE is primarily linked 3 3 8 to auxin activity. Since stress has also been identified as an important factor in SE induction (Mantiri , our findings suggest that at least 3 4 0 in our SE system stress is downstream of auxin signaling, and not a parallel pathway that is 3 4 1 additionally triggered by 2,4-D or its analogues. It is therefore unlikely that a chemical biology 3 4 2 approach as presented here will allow to tease apart the stress and auxin pathways by identifying 3 4 3 compounds that trigger each pathway separately. The new 2,4-D analogues identified here can be 3 4 4 useful tools to study the importance of aspects of auxin physiology either during SE induction or 3 4 5 through their effects on the root system architecture. For example, it would be interesting to 3 4 6 understand why MCPA, which is classified as very active auxin based on root growth inhibition and 3 4 7 auxin response reporter expression, shows a significantly reduced capacity to induce SE. And why 3 4 8 does 3-Cl, classified as weakly active auxin, strongly induce the auxin response in the root 3 4 9 differentiation zone thereby specifically promoting root hair formation? Or how can 3,5-D strongly 3 5 0 inhibit lateral root formation without clearly affecting auxin responsive gene expression? These 3 5 1 differences may lie in the compound specific metabolism or transport characteristics of the 2,4-D 3 5 2 analogues, which has already been shown to differ between the natural auxin IAA and the synthetic Auxin activity of 2,4-D analogues correlates to simulated binding properties to TIR1 3 5 6 The Arabidopsis seedling root growth inhibition assay combined with the use of the pDR5:GUS and 3 5 7 R2D5 auxin response reporters provided consistent results with respect to classifying the 40 2,4-D 3 5 8 orthologs as very active, active or weakly active auxins, or having no auxin activity. Importantly, this 3 5 9 classification was in agreement with previously published results on the 2,4-D analogues PAA, 2,4 3 6 0 Br Molecular dynamics simulations indicated that the auxin activity of 2,4-D analogues could be 3 6 7 correlated to their binding strength to TIR1. One important factor here was the theoretical enthalpy of 3 6 8 the system, as all very active auxin analogues showed an enthalpy value below -12 Kcal/mol. 3 6 9 However, the distance between Pro7 of Aux/IAA and the 2,4-D analogue was also important, as some 3 7 0 of the tested analogues lacking this CH·π interaction showed no or only weak auxin activity, despite 3 7 1 the fact that their enthalpy value was below -12 Kcal/mol.
It is important to note that, due to the expensive cost of entropy calculation combined with low 3 7 3 reliability for such a big system, we did not calculate this parameter. However, we considered that the 3 7 4 molecules are very similar and, although we know there will be differences in entropy among the 3 7 5 derivatives, we assume that they will not change the results dramatically. Based on our screen of 40 compounds, we have identified several trends for the structure auxin 3 7 9 activity relationship of 2,4-D analogues. First, a halogen at the 4-position is more important for auxin 3 8 0 activity than at the 2-position (4-Cl; 2-Cl-4-F; 4-I/Br/F; versus 2-Cl/I/Br/F). Second, the activity of 3 8 1 analogues with a substitution at the 3-or 5-position remains, but such compounds are less active 3 8 2 compared to 2,4-substitutions (2,5-D; 3,4-D; 2,4,5-T). Third, both substitutions at the alpha position 3 8 3 (Mecoprop; 2,4-DP; 2,4-DB; 2,4-DiB) and a longer carboxylate chain with an even number of 3 8 4 carbons are tolerated (MCPB; 2,4-DnB). This finding is in agreement with a previous study that 3 8 5 showed that IAA analogues with a substitution on their carboxylate chain, up to n=4 carbons, can still 3 8 6 bind to TIR1, thus allowing for a modification on this position (Hayashi et al., 2008). Taken these  3  8  7 trends together, we conclude that an electron withdrawing group, in particular a halogen, at the 4-3 8 8 position is important for auxin activity. In addition, we suggest that electron withdrawing or donating 3 8 9 properties of the 2-position are less important for auxin activity and that the size of the 2-substituent 3 9 0 could influence auxin activity, since both methyl and chloro groups are accepted at this position (like 3 9 1 2,4-D and MCPA). There are, however, a few outliers based on this general conclusion: 2,4-F is only 3 9 2 weakly active, but has a strong electron withdrawing group at the 4-position, and 2,4-DnB is also 3 9 3 weakly active, but has the same substitution as 2,4-D and a longer carboxylate chain which does not 3 9 4 disrupt auxin activity for MCPB. Altogether, these results indicate that small modifications at 3 9 5 different parts of 2,4-D can lead to different physiological activities, generating molecules with 3 9 6 interesting other applications in plant (tissue) culture than SE, as described below. 3 9 7 3 9 8 Several 2,4-D analogues are useful tools to modulate the root system architecture 3 9 9 In the assessment of 2,4-D analogues, we have found that some 2,4-D analogues (2,3-D, 3,5-D, 3-Me, 4 0 0 2,4,6-T and 2,4-DnP) either had no effect on or only slightly affected root growth (Fig. 3), whereas 4 0 1 they inhibited the formation of lateral roots (Fig. 6). Since inhibition of lateral root formation is 4 0 2 typical for compounds with anti-auxin activity (Larsen, 2017), we suspected that they may act as anti-4 0 3 auxins. Such anti-auxin activity could be caused by high affinity binding to only one of the co-4 0 4 receptors, thereby preventing the formation of the F-box-Aux/IAA complex. As such, high 4 0 5 concentrations of an anti-auxin would compete effectively with IAA for establishing the TIR1-4 0 6 Aux/IAA complex required for ubiquitination-mediated degradation of the Aux/IAA repressors. 4 0 7 Interestingly, our 2,4-D analogues did not change pDR5 activity, nor did they lead to enhanced or 4 0 8 reduced degradation of DII-VENUS in the root tip, which would be expected for an anti-auxin. 4 0 9 Moreover, the negative effect on lateral root formation varied per compound. For example, 3-Me only 4 1 0 mildly affected lateral root formation, whereas 3,5-D and 2,4-DnP almost completely inhibited this 4 1 1 process, with 3,5-D preferably blocking at stage 2 (enhanced number of stage 2 primordia) whereas 4 1 2 2,4DnP resulted in a reduction of all stages. Clearly, further research on these 2,4-D analogues, such 4 1 3 as a forward genetic screen for mutants that develop lateral roots when grown on the compound, is 4 1 4 required to unravel the exact molecular mechanism by which they repress lateral root formation. As 4 1 5 such, these 2,4-D analogues could uncover new components involved in lateral root formation. At the 4 1 6 same time, they could be used to control lateral root formation, e.g. to prevent branching during early 4 1 7 seedling development to obtain a deeper root system. 4 1 8 Adventitious root (AR) formation is an organogenesis process by which new roots are produced from 4 1 9 non-root tissues. AR induction is an important but often rate limiting step in the vegetative 4 2 0 propagation of many horticultural and forestry plant species. Generally, AR induction is promoted by 4 2 1 auxins, and IBA and NAA are the most common auxins used for this purpose in commercial 4 2 2 operations (Geiss et al., 2009). Our results show that 2,5-D and 3-Cl efficiently induce AR from 4 2 3 Arabidopsis hypocotyls even more efficiently than IBA or NAA. Therefore, 2,5-D and 3-Cl are 4 2 4 recommended as new AR inducers and they may be used to resolve rooting in recalcitrant species. 4 2 5 Roots hairs extend from root epidermal cells in the differentiation zone of the root and support plants 4 2 6 in nutrient absorption, anchorage, and microbe interactions (Lee and Cho, 2013). Exogenous auxin 4 2 7 treatments generally promotes root hair induction and development (Lee and Cho, 2013). We showed 4 2 8 that some 2,4-D analogues (3-Cl, 3,4-D, 2,4-DP, and Mecoprop) induced ectopic root hair formation 4 2 9 on Arabidopsis roots, while promoting lateral root formation and having a relatively mild effect on 4 3 0 root growth compared to 2,4-D itself. Specifically, roots of plants grown on the analogue 3-Cl 4 3 1 produced several times more root hairs compared to other analogues. Unlike other 2,4-D analogues, 3-4 3 2 Cl specifically induced a strong auxin response in the differentiation zone of the root and a weaker 4 3 3 response in the root elongation zone (Fig. S4), this could explain the differential effect of 3-Cl on root 4 3 4 biomass. We recommend 3-Cl as a promising compound for use in horticulture or agriculture to 4 3 5 enhance root hair formation and thereby improve crop performance by enhanced ion and water 4 3 6 uptake. The 2,4-D analogues that were used in this study were divided into 2 main categories. The first one 4 4 2 contained 2,4-D analogues with substituents at different positions on the aromatic ring (Fig. S1). The were plated on ½ MS medium and grown in the tissue culture room at 21°C, 16 hours photoperiod 4 5 5 and 50 % relative humidity. After around 2 weeks, the germinated seeds were planted in soil in the 4 5 6 climate room at 20°C, 16 hours photoperiod and 70 % relative humidity. 4 5 7 4 5 8 Arabidopsis primary root growth assay adventitious root induction 4 5 9 The Arabidopsis seeds were germinated on ½ MS medium. Five-day-old seedlings were transferred 4 6 0 to new ½ MS medium supplemented with 2,4-D analogues. The length of the primary root was 4 6 1 quantified after 3 days, incubation with 2,4-D analogues. Primary root length and lateral root numbers 4 6 2 were analyzed with ImageJ software. 4 6 3 For the adventitious root (AR) induction, the seed was first grown in complete darkness. Six-day-old 4 6 4 seedlings were transferred to new ½ MS medium supplemented with 2,4-D analogues 16 hours 4 6 5 photoperiod. The number of ARs induced from hypocotyls was quantified after 7 days. 4 6 6 4 6 7 Histological GUS staining assay 4 6 8 In order to test auxin activity of 2,4-D analogues, the activity pDR5:GUS reporter was investigated in 4 6 9 the presence of 5 µM 2,4-D analogues. Histochemical β -glucuronidase (GUS) staining was performed 4 7 0 as described previously (Anandalakshmi et al., 1998) with some modifications. Samples were 4 7 1 submerged in 1-2mL staining solution and incubated for 4 hours at 37°C followed by rehydration 4 7 2 through a graded ethanol series 75% -50% -25% for 10 minutes each, with 5 minutes incubation 4 7 3 between each step. Arabidopsis Col-0 on MS medium without any supplement was used as a control. 4 7 4 The tissue-specific GUS staining intensity was quantified as mean grey values by analyzing images of 4 7 5 independent samples capturing the same region of interest (ROI) using ImageJ, as previously For SE induction, 11 days old siliques of Arabidopsis Columbia-0 wild-type were sterilized in 10 % 4 8 0 (v/v) sodium hypochlorite for 10 minutes and then washed with MQ water for 4 times. Sterilized 4 8 1 siliques were dissected to acquire IZEs. IZEs were cultured on B5 medium mixed with 2,4-D 4 8 2 analogues for 14 days in the culture room at 21°C, 16 hours photoperiod and 50 % relative humidity. 4 8 3 The IZEs were then transferred to ½ MS medium for 7 days and the number of SEs was counted Molecular dynamics simulations 4 9 7 Molecular dynamics simulations were run using the crystal structure of the systems comprising the 4 9 8 TIR1, auxin responsive protein fragment, co-factor inositol hexakisphosphate, and the auxin. In and, the second stage minimizes of all the atoms in the simulation cell. Heating the system is the third 5 0 9 step raising gradually the temperature 0 to 300 K under a constant volume and periodic boundary 5 1 0 conditions. In addition, Harmonic restraints of 10 kcal·mol -1 were applied to the solute, and the 5 1 1 Berendsen temperature coupling scheme (Berendsen et al., 1984) was used to control and equalize the 5 1 2 temperature. The time step was kept at 2 fs during the heating phase. Long-range electrostatic effects 5 1 3 were modelled using the particle-mesh-Ewald method (Darden et al., 1993). The Lennard-Jones 5 1 4 interactions cut-off was set a t 8 A . An equilibration step for 2 ns with a 2 fs time step at a constant 5 1 5 pressureand temperature of 300 K was performed prior the production stage. The trajectory 5 1 6 production stage kept the equilibration step conditions and was prolonged for 200 ns longer at 1 fs 5 1 7 time step. Besides, the auxin analogs required a previous preparation step where the parameters and 5 1 8 charges were generated by using the antechamber module of AMBER, using GAFF force field and 5 1 9 AM1-BCC method for charges (Jakalian et al., 2002). In addition, 100 frames evenly separated during 5 2 0 the last 5000 steps of the simulation were employed for the enthalpy calculation by using MM-PBSA 5 2 1 approximation. Supplementary data 5 2 6 The following supplementary data are available online. Chemical synthesis of 2,4-D analogues. 5 3 5 Copies of NMR spectra. All data supporting the findings of this study are available within the paper and within the 5 5 1 supplementary data published online. Raw data are available from the corresponding author, (Remko 5 5 2 Offringa), upon request. Novel materials used and described in the paper are available upon request 5 5 3 for non-commercial research purposes. test. In B-D, Arabidopsis seedlings were initially cultured for 6 days on hormone-free medium in the dark, and 6 1 2 subsequently the seedlings were transferred on medium containing the indicated compound.  also Table S2). Several parameters were 6 3 4 calculated from the molecular dynamics simulations: i) the enthalpy binding of the selected compounds and the 6 3 5 standard deviation (std) was determined, ii) the root-mean-square deviation of the TIR1 protein (Protein RMSD) 6 3 6 with a standard deviation (std) was calculated as a measure of system stability during the simulation trajectory, 6 3 7 iii) the distance between the gamma carbon of the proline of the AUX/IAA peptide and the ring mass center of 6 3 8 the compound and the corresponding standard deviation were determined (Distance gamma). At short distances, 6 3 9 a stronger CH·π interaction is assumed. A larger distances (5Å and more) the CH·π contribution is more or less  m  p  o  u  n  d  A  u  x  i  n  a  c  t  i  v  i  t  y  E  n  t  h  a  l  p  y  b  i  n  d  i  n  g  (  K  c  a  l  /  m  o  l  )  s  t  d  P  r  o  t  e  i  n  R  M  S  D  (   Å   )  s  t  d  D  i  s  t  a  n  c  e  g  a  m  m  a  (   Å   )  s  t  d   2  ,  4  -D  v  e  r  y  a  c  t  i  v  e  -2  6  ,  9  0  ,  8  4  1  ,  4  0  ,  1  2  4  ,  2  0  ,  3  8  I  A  A  v  e  r  y  a  c  t  i  v  e  -2  6  ,  8  0  ,  8  0  1  ,  2  0  ,  1  3  4  ,  0  0  ,  2  8  N  A  A  v  e  r  y  a  c  t  i  v  e  -2  9  ,  6  1  ,  0  6  1  ,  6  0  ,  1  5  4  ,  3  0  ,  3  9  2  ,  4  -D  B  (   S   )  v  e  r  y  a  c  t  i  v  e  -3  0  ,  8  0  ,  7  0  1  ,  3  0  ,  1  6  4  ,  3  0  ,  3  3