Soil microorganisms buffer the reduction in plant growth and physiological performance under combined abiotic stress in the halophyte Salicornia ramosissima

Highlights  The functional responses of plants to complex environmental matrices remain unknown.  This knowledge gap is even greater for stress-tolerant species, such as halophytes.  Extreme combinations of environmental stressors impair their optimal function.  Biological interaction with PGP rhizobacteria ameliorates that negative impact.  The observed beneficial effect is accentuated under atmospheric CO2 enrichment. Highlights 1 Soil microorganisms buffer the reduction in plant growth and physiological 1 performance under combined abiotic stress in the halophyte Salicornia 2 ramosissima 3 4 5 Enrique Mateos-Naranjo1, Jesús Alberto Pérez-Romero2, Giacomo Puglielli1, Javier 6 López-Jurado1,3, Jennifer Mesa-Marín1, Eloísa Pajuelo4, Ignacio David Rodríguez7 Llorente4, Susana Redondo-Gómez1 8 9 1Departamento de Biología Vegetal y Ecología, Facultad de Biología, Universidad de 10 Sevilla, 1095, 41080, Sevilla, Spain 11 2Departamento de Biología, Instituto Universitario de Investigación Marina (INMAR), 12 Universidad de Cádiz, Puerto Real, 11510, Spain 13 3School of Natural Sciences, University of Tasmania, Private Bag 55, Hobart, TAS 14 7001, Australia 15 4Departamento de Microbiologia, Facultad de Farmacia, Universidad de Sevilla, C/ 16 Profesor García González 2, 41012 Sevilla, Spain 17 18 Author for correspondence: 19 Enrique Mateos-Naranjo 20 E-mail: [email protected]s 21 22 REVISED Manuscript (text UNmarked) Click here to view linked References 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 2 ABSTRACT 23 24 The impact of multifactorial abiotic stress combinations on plant functional responses 25 remains controversial, and general patterns of response are yet to emerge. This 26 knowledge gap is particularly relevant for species with innate tolerance to 27 environmental stress. Using the halophyte Salicornia ramosissima as a model species, 28 we performed a multifactorial study with 16 experimental scenarios that included or not 29 beneficial microorganisms in order to quantify their impact on plant growth, 30 photosynthetic performance, osmotic adjustment and ion homeostasis. The experimental 31 scenarios were characterized by the combination of four factors with two levels 32 (salinity: 171 and 510 mM NaCl; water stress: yes and no; temperature min/max range: 33 14/25 and 18/29ºC and atmospheric CO2 concentration: 400 and 700 ppm). A plant 34 growth-promoting rhizobacteria (PGPR) consortium was used as a proxy for positive 35 biological interaction. The results revealed that the multifactorial stress combinations 36 triggered unique functional responses, depending on the stress factors involved. 37 However, there was an overall more negative impact on plant functional traits under the 38 most extreme scenario (i.e., 510 mM NaCl + water stress + high temperature). 39 Interestingly, the presence of PGPR was able to reverse this negative influence, 40 although this effect was negligible under non-stressful conditions. Furthermore, the 41 positive effect of PGPR was even magnified when coexisting with elevated atmospheric 42 CO2 concentration. This response is associated with mitigation of the negative impacts 43 of suboptimal factor combinations on plant growth, photosynthetic 44 performance/efficiency, and water/nutrient homeostasis. Therefore, we conclude that 45 the positive impact of microorganisms on halophyte tolerance in complex 46 environmental matrices would only be determinant under extreme conditions in which 47 plant intrinsic tolerance mechanisms would not be sufficient. Remarkably, this effect 48 could be accentuated by increasing atmospheric CO2 concentration. 49 50 51 Keywords: CO2 enrichment, drought, halophyte, functional traits, multifactorial stress 52 combination, rhizomicrobiome, NaCl-stress, temperature. 53 54 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 3 Highlights 55  The functional responses of plants to complex environmental matrices remain 56 unknown. 57  This knowledge gap is even greater for stress-tolerant species, such as 58 halophytes. 59  Extreme combinations of environmental stressors impair their optimal function. 60  Biological interaction with PGP rhizobacteria ameliorates that negative impact. 61  The observed beneficial effect is accentuated under atmospheric CO2 62 enrichment. 63 64 65 66 67 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 4 1. Introduction 68 In recent decades, stress tolerance in plants has received particular attention from the 69 scientific community as a way to understand: (i) the evolution of plant species and 70 communities in past, current and future climates, and (ii) the mechanisms that provide 71 tolerance, in order to develop crop breeding programmes for increasingly stressful 72 growing environments (Ainsworth et al., 2016). Despite the enormous progress made, 73 many of these studies have been mostly confined to assessing the effect of a few 74 stressors (those characterizing an ecosystem) without considering interactions among 75 multiple environmental constraints (Porter et al., 2020; Puglielli et al., 2021; Zandalinas 76 et al., 2021, 2022a,b). These circumstances have led to an increasing interest in plant 77 responses to multiple concurrent environmental stressors. Thus, many studies have 78 explored the responses of plants to different combinations of abiotic stresses including 79 drought, salinity, temperature, high irradiation, flooding, etc. (Mittler, 2006; Atkinson & 80 Urwin, 2012; Rivero et al., 2013; Zandalinas et al., 2021, 2022a,b). In fact, some effort 81 has been already put into summarizing the main conclusions drawn from those studies. 82 Current evidence indicates that plant responses are unique to each multifactorial stress 83 combination, and positive, negative, and even neutral responses have been described 84 (see Suzuki et al., 2014; Zandalinas et al., 2021). 85 The impact of multifactorial abiotic stress combinations on plant performance 86 depends on the equilibrium between species adaptative and acclimatation limits to stress 87 tolerance, extrinsic factors related to the characteristics of the stressor (i.e., intensity, 88 duration, frequency or its combination; Choudhury et al., 2017; Zandalinas et al., 2021, 89 2022a,b), and the potential coexistence of ecosystem drivers that could modulate the 90 influence of abiotic factors (Suzuki et al., 2014; Rilling et al., 2021; Rivero et al., 2022). 91 In this sense, several works have identified beneficial modulating effects of potential 92 environmental drivers, such as increased plant tolerance to suboptimal salinity 93 concentrations, drought or pollutants due to atmospheric CO2 concentration enrichment 94 (Hibbs et al., 1995; Mateos-Naranjo et al., 2010a,b; Pérez-Romero et al., 2018a,b 95 2019a) or due to the presence of beneficial microorganisms (Mateos-Naranjo et al., 96 2015, 2020; Valle-Romero et al., 2023). However, most of the modulation effects of the 97 described drivers have been settled in simple experimental frameworks often only 98 characterized by pairwise interactions (i.e., abiotic factor vs. driver). Therefore, there is 99 a challenge to determine the functional responses of plants in complex environmental 100 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 5 matrices characterized by the coexistence of several abiotic factors and potential drivers. 101 This information becomes even more important for plants that have evolved the ability 102 to grow in more extreme and potentially stressful environments, such as halophytes 103 (Stavridou et al., 2019; Hussain et al., 2020; Lu et al., 2021; Orrego et al., 2020; 104 Mateos-Naranjo et al., 2021). However, only a few studies have investigated the 105 responses of these species to combinations of more than two abiotic stressors , and none 106 of them considered the role of interacting environmental drivers of different nature in 107 triggering plant responses to these multifactorial scenarios. 108 In this paper, we tested the influence of a complex environmental matrix 109 (characterized by different abiotic factors occurring simultaneously; Fig. 1) on key 110 functional traits of the C3 halophyte Salicornia ramosissima J. Woods 111 (Chenopodiaceae). In addition, we evaluated how environmental factors, namely 112 different atmospheric CO2 concentrations and the presence/absence of beneficial soil 113 microorganisms, modulate plant responses to the multi-stress setup. For that purpose, 114 plant growth, photosynthetic performance, osmotic adjustments, and ion homeostasis 115 were explored. S. ramosissima represents a suitable model species, since it has stress 116 tolerance mechanisms that are shared by most halophytes, and inhabits salt marshes that 117 are exposed to a wide range of environmental constraints (e.g., salinity, drought, 118 temperature extremes; Pérez-Romero et al., 2018a, 2019a). Furthermore, previous 119 studies have evaluated the positive or negative effects of atmospheric CO2 enrichment 120 on its tolerance to environmental factors (Pérez-Romero et al., 2018a,b; 2019b). 121 Recently, Mesa-Marín et al. (2019) found and characterized several bacterial strains 122 with plant growth-promoting properties (PGP) that are compatible with S. ramosissima. 123 This allows us to assess the effect of biological interactions on plant tolerance 124 mechanisms under a complex environmental matrix. 125 We hypothesized that plant functional responses to multifactorial stress combination 126 are modulated by coexisting environmental drivers, which would be linked to the 127 activation of concrete halophytic tolerance mechanisms. Thus, we address the following 128 specific questions: (i) Would the impact of the multifactorial stress combination (i.e., 129 salinity x high temperature x drought) on plant performance in a similar way in terms of 130 sign and magnitude when PGP microorganisms are present? (ii) Would this response be 131 modulated by other potential environmental drivers, such as atmospheric CO2 132 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 6 enrichment? (iii) Would these response patterns be mediated by different degrees of 133 activation of specific plant tolerance mechanisms? 134 135 2. Material and methods 136 2.1. Plant material 137 Salicornia ramosissima seeds were harvested from individual plants (n = 30) randomly 138 selected from a well-established population located in Odiel salt marshes (37º15’N, 139 6º58’O; SW Spain) and immediately transported to the laboratory and stored at 4° C (in 140 the dark) for 3 months. After the storage period, the seeds were surface disinfected by 141 immersion and vigorous shaking in 5% sodium hypochlorite (v/v) for 1 min, followed 142 by repeated washes with sterile distilled water. The sterilized seeds were then placed in 143 a seed germination chamber (ASL Aparatos Cientificos M-92004, Madrid, Spain) and 144 subjected to a day/night regime of 16 h of light (photon flux rate, 400–700 nm, 35 145 μmolm−2 s−1) at 25 °C and 8 h of darkness at 12 °C, for 15 days. The germinated 146 seedlings were planted in 0.25 L individual plastic plots containing an organic 147 commercial substrate (Gramoflor GmbH and Co. KG.) and a sand mixture (2:1) 148 previously disinfected, and placed in controlled environment walk-in chambers 149 (Aralab/Fiberoclima 18.000EH, Lisbon, Portugal). These were programmed with an 150 alternate daily regime of 16/8 h and 25/14 ℃ (light/darkness), 300 μmol m−2 s−1 light 151 intensity, 50 ± 5% relative humidity. Plants were watered with a 171 mM NaCl saline 152 solution and soil was maintained fully saturated. 153 The temperature range was chosen as the daily mean maximum and minimum 154 temperature recorded for southwestern Iberian Peninsula during spring (aemet.es), the 155 season when S. ramosissima presents its vegetative phase. Both salinity concentration 156 and soil irrigation level used were optimal for S. ramosissima development, as 157 previously determined (Pérez-Romero et al., 2018a,b; 2019a,b). 158 159 2.2. Experimental design and treatment characteristics 160 A total of 256 plants grew under the previously described conditions until they reached 161 a mean height of 11 cm (approx. after 30 days), at which time the experiment began. 162 Plants were then randomly divided into 16 different experimental scenarios defined by 163 the combination of four different experimental abiotic factors, which were: two 164 concentrations of atmospheric CO2 (400 ppm and 700 ppm CO2), two temperature 165 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 7 ranges (darkness/light: 14/25 ℃ and 18/29 ℃), two salinity concentrations (171 mM and 166 510 mM NaCl) and two irrigation regimes (well-watered, WW, and water-stressed, WS 167 hereafter) (see Table 1 for details on experimental scenarios). And subsequently plants 168 were subjected to two inoculation treatments, which generated a total of 32 treatment 169 combinations (i.e., 16 experimental scenarios x two inoculation treatments; Fig. 1). 170 Lower level values of CO2 concentration, temperature range and salinity were chosen to 171 represent the most favourable natural environmental conditions for S. ramosissima 172 growth. The higher atmospheric CO2 and temperature level values were included to 173 reproduce climate change forecasts by the end of 2100 (i.e., 700 ppm CO2 and an 174 increase in average temperature by 4ºC; IPPC, 2014). The higher salinity concentration 175 (i.e., 510 mM NaCl) and lower irrigation regime were selected to reproduce temporary 176 hypersaline conditions and drought registered at the Gulf of Cadiz inland salty habitat, 177 where natural populations of S. ramosissima occur (Pérez-Romero et al., 2019a). 178 Furthermore, this set of scenarios was divided into two blocks with different inoculation 179 treatments (either without inoculum or with bacterial inoculation). 180 For the establishment of experimental conditions for atmospheric CO2 concentration 181 and temperature, plants were placed in four different controlled environment chambers, 182 which were programmed with the specific atmospheric CO2 concentration and the 183 minimum/maximum temperature range (i.e., 400 ppm CO2 + 14/25 ℃, 400 ppm CO2 + 184 18/29 ℃, 700 ppm CO2 + 14/25 ℃ and 700 ppm CO2 + 18/29 ℃). Atmospheric CO2 185 concentrations in chambers were continuously monitored by CO2 sensors and 186 maintained by supplying pure CO2 from a compressed gas cylinder (Air Liquide, B50 187 35K). For the establishment of saline and irrigation treatments inside each chamber, all 188 plants were watered with the specific saline concentration (i.e., 171 or 510 mM NaCl) to 189 field capacity at the beginning of the experiment. Then, they were divided into four 190 groups of treatments and arranged in a randomized plot to reach the combination of 191 salinity and irrigation regimes. Two groups were kept under WW conditions throughout 192 the experimental period by placing pots in individual plastic trays containing the 193 appropriate NaCl solution to a depth of 1 cm, thus allowing well-watered irrigation for 194 both experimental salinity levels (i.e., 171 mM NaCl + WW and 510 mM NaCl + WW, 195 respectively). To avoid changes in NaCl concentration caused by water evaporation, 196 water levels in the trays were continuously monitored throughout the experimental 197 period. Individuals from the other two groups were deprived of watering (WS). In this 198 treatment, irrigation was suppressed for 5 days before the end of the experiment. At that 199 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 8 time, soil water content had dropped by 40% based on the difference between the mass 200 of WS pots and that of WW pots, thus obtaining water stress treatments for both salinity 201 levels (i.e., 171 mM NaCl + WS and 510 mM NaCl + WS, respectively). 202 Finally, in terms of the PGP treatment, plants were inoculated twice: one day after 203 the imposition of the experimental treatments, except for drought, and in the middle of 204 the experimental period. The bacterial inoculum used in this work was formed by 205 strains Bacillus methylotrophicus SMT38, Bacillus aryabhattai SMT48, and Bacillus 206 licheniformis SMT51. These rhizobacteria were originally isolated from halophyte 207 rhizospheres, including S. ramosissima, that commonly inhabit salt marshes in 208 southwestern Spain (Mesa-Marín et al., 2019). These bacteria showed high resistance to 209 environmental stress, such as salinity (up to 2 M NaCl) and temperature (up to 40ºC) 210 and presented multiple PGP properties, such as nitrogen fixation, phosphate 211 solubilization, biofilm formation capacity, siderophores production, indole-3-acetic 212 acid, and ACC deaminase (Mesa-Marín et al., 2019). For technical details of inoculum 213 preparation and application, see Mesa-Marín et al. (2018). 214 215 2.3. Plant growth analysis 216 At the end of the experiment, all plants from each treatment combination were harvested 217 and divided into branches and roots. These biomass fractions were then oven dried at 60 218 ºC for 48 h and weighed to obtain estimates of the dry mass content of the leaf (LDMC) 219 and root (RDMC). Additionally, before and after the treatments, the height of the main 220 branches (plant height, Ph) was measured. 221 2.4. Plant water status analysis 222 The osmotic potential, Ψo, and the water content, BWC, were measured in fully 223 developed random primary branches (n = 5 per treatment) at the end of the experiment. 224 Ψo was determined using the psychrometric technique with a vapor pressure osmometer 225 (5600 Vapro, Wescor, Logan, USA). BWC was calculated as follows: 226 BWC (%) = [(FW - DW) / (DW)] x 100 227 Where FW = fresh weight of the branches and DW = dry weight after drying in an oven 228 at 60 °C for 48 h. 229 230 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 15 of microorganisms, especially plant growth-promoting bacteria (PGPB) (Mayak et al., 423 2004; Wang et al., 2016; Chen et al., 2016; Forni et al., 2017; Navarro-Torre et al., 424 2017; Paredes-Páliz et al., 2018; Backer et al., 2018; Mateos-Naranjo et al., 2015, 425 2021), and atmospheric CO2 enrichment on plant growth and physiological performance 426 under stress conditions (Pérez-Romero et al., 2018, 2019b; Mateos-Naranjo et al., 427 2010b, 2021). However, this is, to the best of our knowledge, the first study considering 428 a complex multifactorial abiotic stress matrix. Importantly, our results revealed 429 increased plant growth, especially at the root level, under the most extreme conditions 430 in inoculated plants at elevated atmospheric CO2 concentration compared to non431 inoculated individuals. This response indicates that the presence of microorganisms 432 would contribute to the maintenance of halophyte growth under extreme conditions and 433 that the importance of this modulation effect is accentuated by atmospheric CO2 434 enrichment. An alternative and/or complementary interpretation is that the multifactorial 435 abiotic stress combinations affected plant–microbiome interactions. According to the 436 postulates of Rilling et al. (2021), environmental factors could be determining factors 437 for the generation of non-additive functional responses by impacting groups of soil 438 biota that interacts with plants. In addition to generating direct multifactorial effects on 439 plant functional traits as previously indicated, stress factors could have led to changes 440 within the microbiological component and consequently to its functional characteristics 441 that modulate PGP properties (Jansson & Hofmockel, 2020; Marín & Kohout, 2021; 442 Yang et al., 2021). We encourage future research to include direct measurements of 443 functional responses of the microbiome at the end of the experiment to confirm this 444 idea. Despite this limitation, the bacterial consortium used in this study was integrated 445 by strains with multiple plant growth promoting properties and environmental resistance 446 of up to 2 M NaCl and 40ºC (Mesa-Marín et al., 2019), as already mentioned. 447 Therefore, it would be reasonable to expect a positive effect of the full factorial 448 combination on the microbiological component and in consequence on plant 449 performance. This was indeed reflected by synergistic positive responses mainly at 450 growth level, and by the maintenance of water content and a better osmotic adjustment 451 under the most extreme multifactor combinations, as described below. 452 Our results also suggest that the modulating effect of bacterial inoculation on plant 453 stress tolerance was partially mediated by an overall protection of several key steps in 454 the photosynthetic pathway, mainly associated with non-stomatal traits (Flexas et al., 455 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 16 2002; López-Jurado et al., 2020; Pan et al., 2021). Although higher AN values were 456 recorded in inoculated plants grown under extreme environmental matrix and elevated 457 atmospheric CO2 concentration, this response was not associated with variations in gs 458 between the two inoculation treatments. Thus, differences in carbon assimilation 459 between inoculation treatments are likely to be related to an ameliorative effect exerted 460 by the bacteria on biochemical limitations imposed by abiotic stress combinations. This 461 agrees with a better functionality of the RuBisCO carboxylation capacity (Cen & Sage, 462 2005; Pérez-Romero et al., 2018a, 2019a), as indicated by the lowest Ci values recorded 463 in inoculated plants. Similarly, there was a synergistic potential driver impact on the 464 functionality of the photochemical apparatus, as indicated by the higher Fv/Fm registered 465 in inoculated individuals. This suggests that the combination of beneficial 466 microorganisms and elevated atmospheric CO2 concentration would synergistically 467 contribute to reduce plant photoinhibition (Werner et al., 2002). Consequently, these 468 effects would lead to a reduced limitation in carbon assimilation and captured energy 469 transformation in plant photosystems, a circumstance that would affect photosynthetic 470 productivity and growth (Flexas et al., 2002). This explanation also supports the greater 471 growth measured in inoculated plants under the most extreme conditions. 472 Improved plant water relations were another positive physiological response 473 associated with bacterial inoculation and atmospheric CO2 enrichment. In this regard, 474 the ability of the plant to maintain the optimal tissue water content thanks to its 475 osmoregulatory capacity when exposed to high levels of environmental stress represents 476 an important adaptation (Touchette et al., 2006; Pérez-Romero et al., 2020). Our results 477 revealed that the most extreme environmental factor combination imposed the greatest 478 impact on plant water status, measured as iWUE and WC. Thus, low iWUE and WC led 479 to the activation of the osmoregulatory capacity of the plant (as indicated by the most 480 negative Ψo values), which is essential to prevent water losses (Hormaetxe et al., 2006; 481 Cui et al., 2020). However, we found a down-regulation of the osmotic adjustment in 482 inoculated plants, which would be indicative of a lower level of stress and can be 483 partially explained by their increased root-to-shoot ratio, by means of increased RDMC 484 in inoculated plants of S. ramosissima. Therefore, the direct PGP effect of inoculated 485 strains, such as IAA synthesis and ACC activity, which have been shown to promote 486 plant growth and alleviate stress (Glick, 2012; Mesa-Marín et al., 2019), would 487 contribute to explain these positive results. Additionally, this greater root development 488 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 17 would improve the ability of the plant to absorb nutrients and water, contributing to its 489 more optimal water and nutritional balance, as also corroborated the higher 490 concentrations of Ca and K in the roots and Na in the branches. In this regard, it is 491 possible that alternative stress tolerance mechanisms not evaluated here are contributing 492 to the improvement of water relations in S. ramosissima under these experimental 493 conditions, via accumulation of osmolytes, phytohormonal responses, or antioxidative 494 machinery modulation (Flowers & Colmer, 2008, 2015; Flower et al., 2010). This area 495 is therefore worthy of further future research. 496 In summary, our findings demonstrated that stress-tolerant plants, such as the 497 halophyte S. ramosissima, would experience a negative impact on their main functional 498 traits by a multifactorial combination of abiotic factors. In turn, this could be modulated 499 by other potential environmental drivers, which would be able to reverse the negative 500 impact imposed by stress factors through the activation of more efficient stress tolerance 501 responses. We demonstrated here that those responses contribute to the maintenance of 502 growth and water relations, as well as the functionality of photosynthetic metabolism. 503 Therefore, we can conclude that the beneficial impact of microorganisms on the 504 tolerance of halophytes would only be determinant when their intrinsic tolerance 505 mechanisms are exceeded, meaning under extreme abiotic stress. This contrasts with the 506 general pattern observed in stress-sensitive species, for which environmental drivers 507 (such as positive interactions with PGP bacteria) would enhance plant performance even 508 under benign conditions. Furthermore, this positive modulation effect could be 509 accentuated in a future scenario of even higher atmospheric CO2 concentration, that can 510 compromise the development of S. ramosissima and its capacity for environmental 511 adaptation. 512 513 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 18 Acknowledgements 514 This work has been supported by the Spanish Government (Ministerio de Ciencia e 515 Innovación-AEI) through the grant project PID2021-124750NB-I00, MCIN/ AEI 516 /10.13039/501100011033/FEDER, una manera de hacer Europa, and TED2021517 131605B-I00, MCIN/AEI /10.13039/501100011033 and by the Unión Europea 518 NextGenerationEU/ PRTR and by the grants IJC2020-043331-I funded by MCIN/AEI 519 /10.13039/501100011033, and PID2021-122214NA-I00 funded by MCIN/AEI/ 520 10.13039/501100011033 and by FEDER “ESF Investing in your future”. We also like 521 to thank the University of Seville Greenhouse and Herbarium Research General 522 Services (SGI and SGH, CITIUS) for providing facilities and equipment. We also 523 appreciate the valuable feedback from reviewers insofar as it has helped us improve and 524 enrich our work. 525 526 Conflict of interest 527 The authors declare that they have no conflict of interest. 528 529 CRediT authorship contribution statemen 530 E.M.-N. and. S.R.-G. conceived the study, supervised and acquired funding for the 531 project, E.M.-N. gathered the data, designed and performed the analyses with the help 532 of G.P. and J.L.-J. E.M.-N wrote the first draft of the manuscript with the inputs of S.R.- 533 G. and G.P. J.M.-M. with assistance from E.P. and I.D.R.-L. provided the bacterial 534 inocula. All authors performed the experimental development, provided corrections to 535 manuscript drafts, and discussed ideas within it. 536 537 Declaration of Competing Interest 538 The authors declare that they have no known competing financial interests or 539 personal relationships that could have appeared to influence the work reported in this 540 paper. 541 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 19 References 542 Ainsworth, E.A., Bernacchi, C.J., & Dohleman, F.G. (2016). Focus on ecophysiology. 543 Plant Physiology 172, 619-621. https://doi.org/10.1104/pp.16.01408 544 Alam, H., Khattak, J.Z.K., Ksiksi, T.S., Saleem, M.H., Fahad, S., Sohail, H., Ali., 545 …Alkahtani, J. (2021). Negative impact of long-term exposure of salinity and 546 drought stress on native Tetraena mandavillei L. Physiologia Plantarum 2, 1336547 1351. https://doi.org/10.1111/ppl.13273 548 Atkinson, N.J. & Urwin, P.E. (2012). The interaction of plant biotic and abiotic stresses: 549 from genes to the field. Journal of Experimental Botany 63, 3523-3543. 550 https://doi.org/10.1093/jxb/ers100 551 Backer, R., Rokem, J.S., Ilangumaran, G., Lamont, J., Praslickova, D., Ricci, E., … 552 Smith DL. (2018). Plant growth-promoting rhizobacteria: context, mechanisms of 553 action, and roadmap to commercialization of biostimulants for sustainable 554 agriculture. Frontier in Plant Science 9, 1473. 555 https://doi.org/10.3389/fpls.2018.01473 556 Ben Hamed, K., Ellouzi, H., Talbi, O.Z., Hessini, K., Slama, I., Ghnaya, … Abdelly, C. 557 (2013). Physiological responses of halophytes to multiple stresses. Functional Plant 558 Biology 40, 883-896. https://doi.org/10.1071/FP13074. 559 Carmona, C.P., Pavanetto, N., Puglielli G (2023). Creating and Representing Functional 560 Trait Spaces, ‘funspace’ R package. URL https://CRAN.R561 project.org/package=funspace 562 Cen, Y.P., & Sage, R.F. (2005). The regulation of Rubisco activity in response to 563 variation in temperature and atmospheric CO2 partial pressure in sweet potato. Plant 564 physiology 139, 979-990. https://doi.org/10.1104/pp.105.066233 565 Chen, Y.E., Liu, W.J., Su, Y.Q., Cui, J.M., Zhang, Z.W., Yuan, M., …Yuan, S. (2016). 566 Different response of photosystem II to short and long-term drought stress in 567 Arabidopsis thaliana. Physiologia Plantarum 158, 225568 235. https://doi.org/10.1111/ppl.12438 569 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 20 Choudhury, F.K., Rivero, R.M., Blumwald, E., & Mittler, R. (2017). Reactive oxygen 570 species, abiotic stress and stress combination. The Plant Journa 190, 856-867. 571 https://doi.org/10.1111/tpj.13299. 572 Cui, Y.N., Li, X.T., Yuan, J.Z., Wang, F.Z., Guo, H., Xia, Z.R., … Ma, Q. (2020). 573 Chlorides is beneficial for growth of the xerophyte Pugionium cornutum by 574 enhancing osmotic adjustment capacity under salt and drought stresses. Journal of 575 Experimental Botany 71, 4215-4231. https://doi.org/10.1093/jxb/eraa158 576 Flexas, J., Bota, J., Escalona, J.M., Sampol, B., & Medrano H. (2002). Effects of 577 drought on photosynthesis in grapevines under field conditions: an evaluation of 578 stomatal and mesophyll limitations. Functional Plant Biolology 29, 461–471. 579 https://doi.org/10.1071/PP01119. 580 Flowers, T.J., & Colmer, T.D. (2008). Salinity tolerance in halophytes. New Phytologist 581 179, 945-963. https://doi.org/10.1111/j.1469-8137.2008.02531.x 582 Flowers, T.F., & Colmer, T.D. (2015). Plant salt tolerance: adaptations in halophytes. 583 Annals of Botany 115, 327-331. https://doi.org/10.1093/aob/mcu267 584 Flowers, T.J., Galal, H.K., & Bromham, L. (2010). Evolution of halophytes: multiple 585 origins of salt tolerance in land plants. Functional Plant Biology 37, 604-612. 586 https://doi.org/10.1071/FP09269 587 Forni, C., Duca, D., & Glick, B.R. (2017). Mechanisms of plant response to salt and 588 drought stress and their alteration by rhizobacteria. Plant and Soil 410, 335-356. 589 https://doi.org/10.1007/s.11104-016-3007-x 590 Glick, B.R. (2012). Plant growth-promoting bacteria: mechanisms and applications. 591 Scientifica 15, 963401. https://doi.org/10.6064/2012/963401 592 Hibbs, D.E., Chan, S.S., Castellano, M., Niu, C.H. (1995). Response of red alder 593 seedling to CO2 enrichment and water-stress. New phytologist 129, 569-577. 594 https://doi.org/10.1111/j.1469-8137.1995.tb03024.x 595 Hormaetxe, K., Becerril, J.M., Hernández, A., Esteban, R., & García-Plazaola, J.I. 596 (2006). Plasticity of photoprotective mechanisms of Buxus sempervirens L. Leaves 597 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 21 in response to extreme temperatures. Plant Biology 9, 59-68. 598 https://doi.org/10.1055/s-2006-924456 599 Hussain, T., Koyro, H.W., Zhang, W., Liu, X., Gul, B., & Liu, X. (2020). Low salinity 600 improves photosynthetic performance in Panicum antidotable under drought stress. 601 Frontiers in Plant Sciences 11, 481. https://doi.org/10.3389/fpls.2020.00481 602 IPCC (2014). Climate Change 2014: Impacts, Adaptation, and Vulnerability. In Part A: 603 Global and Sectoral Aspects. Contribution of Working Group II to the Fifth 604 Assessment Report of the Intergovernmental Panel on Climate Change; Cambridge 605 University Press: Cambridge, UK; New York, NY, USA. 606 Jansson, J.K., Hofmockel, K.S. (2020). Soil microbiomes and climate change. Nature 607 Reviews Microbiology 18, 35–46. https://doi.org/10.1038/s41579-019-0265-7 608 López-Jurado, J., Balao, F., & Mateos-Naranjo, E. (2020). Polyploidy-mediated 609 divergent light-harvesting and photoprotection strategies under temperature stress 610 in a Mediterranean carnation complex. Environmental and Experimental Botany 611 171, 103956. https://doi.org/10.1016/j.envexpbot.2019.103956 612 Lu, Y., Zhang, B., Li, L., Zeng, F., & Li, X. (2021). Negative effects of long-term 613 exposure to salinity, drought, and combined stresses on halophyte Halogeton 614 glomeratus. Physiologia Plantarum 173, 2307-2322. 615 https://doi.org/10.1111/ppl.13581 616 Marín, C., & Kohout, P. (2021). Response of soil fungal ecological guilds to global 617 changes. New Phytologist 229, 656-658. https://doi.org/10.1111/nph.17054 618 Mateos-Naranjo, E., Redondo-Gómez, S., Andrades-Moreno, L., & Davy, A.J. (2010a). 619 Growth and photosynthetic responses of the cordgrass Spartina maritima to CO2 620 enrichment and salinity. Chemosphere 81, 725-731. 621 https://doi.org/10.1016/j.chemosphere.2010.07.047 622 Mateos-Naranjo, E., Redondo-Gómez, S., Álvarez, R., Cambrollé, J., Gandullo, J., & 623 Figueroa, M.E. (2010b). Synergic effect of salinity and CO2 enrichment on growth 624 and photosynthetic responses of the invasive cordgrass Spartina densiflora. Journal 625 of Experimental Botany 61, 1643-1654. https://doi.org/10.1093/jxb/erq029 626 Mateos-Naranjo, E., Andrades-Moreno, L., & Redondo-Gómez, S. (2012). Tolerance to 627 and accumulation of arsenic in the cordgrass Spartina densiflora Brongn. 628 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 22 Bioresource Technology 104,187-194. 629 https://doi.org/10.1016/j.biortech.2011.11.006 630 Mateos-Naranjo, E., Mesa, J., Pajuelo, E., Pérez-Martín, A., Caviedes, M.A., & 631 Rodríguez-Llorente, I.D. (2015). Deciphering the role of plant growth-promoting 632 rhizobacteria in the tolerance of the invasive cordgrass Spartina densiflora to 633 physicochemical properties of salt-marsh soils. Plant and Soil 394, 45-55. 634 https://doi.org/10.1007/s11104-015-2504-7 635 Mateos-Naranjo, E., López-Jurado, J., Redondo-Gómez, S., Pérez-Romero, J.A., Glick, 636 B.R., Rodríguez-Llorente, I.D., … Mesa-Marín, J. (2020). Uncovering PGPB 637 Vibrio spartinae inoculation-triggered physiological mechanisms involved in the 638 tolerance of Halimione portulacoides to NaCl excess. Plant Physiology and 639 Biochemistry 154,151-159. https://doi.org/10.1016/j.plaphy.2020.05.034 640 Mateos-Naranjo, E., López-Jurado, J., Mesa-Marín, J., Luque, C.J., Castellanos, E.M., 641 Pérez-Romero, J.A., & Redondo-Gómez, S. (2021). Understanding the impact of a 642 complex environmental matrix associated with climate change on the European 643 marshes engineer species Spartina martima. Environmental and Experimental 644 Botany 182, 104304. https://doi.org/10.1016/j.envexpbot.2020.104304 645 Mayak, S., Tirosh, T., & Glick, B.R. (2004). Plant growth-promoting bacteria that 646 confer resistance in tomato to salt stress. Plant Physiology and Biochemistry 42, 647 565-572. https://doi.org/10.1016/j.plaphy.2004.05.009 648 Mesa-Marín, J., Del-Saz, N., Rodríguez-Llorente, I.D., Redondo-Gómez, S., Pajuelo, 649 E., Ribas-Carbó, M., & Mateos-Naranjo, E. (2018). PGPR reduce root respiration 650 and oxidative stress enhancing Spartina maritima root growth and heavy metal 651 rhizoaccumulation. Frontiers in Plant Science 871, 1500. 652 https://doi.org/10.3389/fpls.2018.01500 653 Mesa-Marín, J., Pérez-Romero, J.A., Mateos-Naranjo, E., Bernabeu-Meana, M., 654 Pajuelo, E., Rodríguez-Llorente, I.D., & Redondo-Gómez, S. (2019). Effect of plant 655 growth-promoting rhizobacteria on Salicornia ramosissima seed germination under 656 salinity, CO2 and temperature stress. Agronomy 9, 655. 657 https://doi.org/10.3390/agronomy9100655 658 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 23 Mishra, A., & Tanna, B. (2017). Halophytes: potential resources for salt stress tolerance 659 genes and promoters. Frontiers in Plant Sciences 8, 829. 660 https://doi.org/10.3389/fpls.2017.00829 661 Mittler, R. (2006). Abiotic stress, the field environment and stress combination. Trends 662 in Plant Science 11, 15-19. https://doi.org/10.1016/j.tplants.2005.11.002 663 Navarro-Torre, S., Barcia-Piedra, J.M., Mateos-Naranjo, E., Redondo-Gómez, S., 664 Camacho, M., Caviedes, M.A., … Rodríguez-Llorente, I.D. (2017). Assessing the 665 role of endophytic bacteria in the halophyte Arthrocnemum macrostachyum salt 666 tolerance. Plant Biology 19, 249-256. https://doi.org/10.1111/plb.12521 667 Orrego, F., Ortiz-Calderon, C., Lutts, S., & Ginocchio, R. (2020). Growth and 668 physiological effects of single and combined Cu, NaCl, and water stresses on 669 Atriplex atacamensis and A. halimus. Environmental and Experimental Botany 670 169,103919. https://doi.org/10.1016/j.envexpbot.2019.103919 671 Pan, T., Liu, M.M., Kreslavski, V.D., Zharmukhamedov, S.K., Nie, C.R., Yu, M., … 672 Shabala, S. (2021). Non-stomatal limitation of photosynthesis by soil salinity. 673 Critical Reviews in Environmental Science and Technology 51, 791-825. 674 https://doi.org/10.1080/10643389.2020.1735231 675 Paredes-Páliz, K., Rodríguez-Vázquez, R., Duarte, B., Caviedes, M.A., Mateos676 Naranjo, E., Redondo-Gómez, S., … Pajuelo, E. (2018). Investigating the 677 mechanisms underlying phytoprotection by plant growth-promoting rhizobacteria 678 in Spartina densiflora under metal stress. Plant Biology 20, 497-506. 679 https://doi.org/10.1111/plb.12693 680 Pérez-López, U., Miranda-Apocada, J., Muñoz-Rueda, A., & Mena-Petite, A. (2013). 681 Lettuce production and antioxidant capacity are differentially modified by salt 682 stress and light intensity under ambient and elevated CO2. Journal of Plant 683 Physiology 170, 1517–1525. https://doi.org/10.1016/j.jplph.2013.06.004 684 Pérez-Romero, J.A., Idaszkin, Y., Barcia-Piedras, J.M., Duarte, B., Redondo-Gomez, 685 S., Caçador, I., & Mateos-Naranjo, E. (2018a). Disentangling the effect of 686 atmospheric CO2 enrichment on the halophyte Salicornia ramosissima J. Woods 687 physiological performance under optimal and suboptimal saline conditions. Plant 688 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 24 Physiology and Biochemistry 127, 617-629. 689 https://doi.org/10.1016/j.plaphy.2018.04.041 690 Pérez-Romero, J.A., Idaszkin, Y., Duarte, B., Baeta, A., Marques, J.C., Redondo691 Gomez, S., … Mateos-Naranjo, E. (2018b). Atmospheric CO2 enrichment effect on 692 the Cu-tolerance of the C4 cordgrass Spartina densiflora. Journal of Plant 693 Physiology 220, 155-166. https://doi.org/10.1016/j.jplph.2017.11.005 694 Pérez-Romero, J.A., Duarte, B., Barcia-Piedras, J.M., Matos, A.R., Redondo-Gómez, 695 S., Caçador, I., & Mateos-Naranjo E. (2019a). Investigating the physiological 696 mechanisms underlying Salicornia ramosissima response to atmospheric CO2 697 enrichment under coexistence of prolonged soil flooding and saline excess. Plant 698 Physiology and Biochemistry 135, 149–159. 699 https://doi.org/10.1016/j.plaphy.2018.12.003 700 Pérez-Romero, J.A., Barcia-Piedras, J.A., Redondo-Gómez, S., & Mateos-Naranjo, E. 701 (2019b). Impact of short-term extreme temperature events on physiological 702 performance of Salicornia ramosissima J. Woods under optimal and sub-optimal 703 saline conditions. Scientific Reports 9, 659. https://doi.org/10.1038/s41598-018704 37346-4 705 Pérez-Romero, J.A., Mateos-Naranjo, E., López-Jurado, J., Redondo-Gómez, S., & 706 Torres-Ruiz, J.M. (2020). Importance of physiological traits vulnerability in 707 determine halophytes tolerance to salinity excess: A comparative assessment in 708 Atriplex halimus. Plants 9, 690. https://doi.org/10.3390/plants9060690 709 Porter, S.S., Bantay, R., Friel, C.A., Garoutte, A., Gdanetz, K., Ibarreta, K,,… Friesen, 710 M.L. (2020). Beneficial microbes ameliorate abiotic and biotic sources of stress on 711 plants. Functional Ecology 34, 2075-2086. https://doi.org/10.1111/1365712 2435.13499 713 Puglielli, G., Hutchings, M.J., & Laanisto, L. (2021). The triangular space of abiotic 714 stress tolerance in woody species: a unified trade‐off model. New Phytologist 229 715 (3), 1354-1362. https://doi: 10.1111/nph.16952 716 R Core Team. (2020). R: a language and environment for statistical computing. R 717 Foundation for Statistical Computing, Vienna, Austria. URL http:// www.R718 project.org/. 719 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 31 abovecompared to belowground organs. Boxplots summarize medians, interquartile 869 range (25th and 75th percentile, box limits) and values greater than interquartile range 870 (whisker limits). Open symbols are values exceeding the interquartile range. *** 871 indicates significant difference at p ≤ 0.05 (two-way ANOVA, statistics and sample size 872 are reported in Supporting Table S1). 873 Fig. S6. Effect size of the difference between inoculated and non-inoculated plants in 874 their position along (a) PC1 and (b) PC2 within each considered scenario obtained after 875 Tukey test. Bars represent 95% confidence interval of the effect size estimate. 876 Differences are considered significant when confidence interval do not include zero. 877 Arrows indicate the two greatest effect sizes (see Results section in the main text). 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 32 Table S1. Results of the two-way ANOVA testing for the effect of inoculation (two 894 levels: Inoculated, Non-inoculated), scenario (16 levels, see material and methods 895 section; Table 1) and their interaction on PC1 and PC2 (see Data analysis section in the 896 main text). Degrees of freedom, F-statistics and p-values are shown. 897 Response Effect d.o.f. F-value p-value PC1 Inoculation 1 47.38 < 0.0001 Scenario 15 86.35 < 0.0001 Inoculation*Scenario 15 2.01 < 0.05 Residuals 121 PC2 Inoculation 1 6.29 < 0.05 Scenario 15 3.10 < 0.001 Inoculation*Scenario 15 1.11 0.35 Residuals 121 898 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 171 mM NaCl WW Irrigation Salinity 700 ppm CO2 Non-Inoculated Inoculated 400 ppm CO2 400 ppm CO2 14/25 ℃ 18/29 ℃ 700 ppm CO2 WS 510 mM NaCl 14/25 ℃ 18/29 ℃ S. ramosissima Figure 1 −15 −10 −5 0 5 10 −5 0 5 PC1 (35.78%) PC2 (21.8%) Data pooled −15 −10 −5 0 5 10 −5 0 5 400 ppm −15 −10 −5 0 5 10 −5 0 5 700 ppm PC1 (35.78%) PC2 (21.8%) Inoculation 1 0 Figure 2 -4 -2 0 2 In N-In PC1 Scenario 1 (a) -4 -2 0 2 In N-In PC1 Scenario 2 (b) -4 -2 0 2 In N-In PC1 Scenario 3 (c) -4 -2 0 2 In N-In PC1 Scenario 4 (d) -4 -2 0 2 In N-In PC1 Scenario 5 (e) -4 -2 0 2 In N-In PC1 Scenario 6 (f) -4 -2 0 2 In N-In PC1 Scenario 7 (g) -4 -2 0 2 In N-In PC1 Scenario 8 (h) -4 -2 0 2 In N-In PC1 Scenario 9 (i) -4 -2 0 2 In N-In PC1 Scenario 10 (k) -4 -2 0 2 In N-In PC1 Scenario 11 (l) -4 -2 0 2 In N-In PC1 Scenario 12 (m) -4 -2 0 2 In N-In PC1 Scenario 13 (n) -4 -2 0 2 In N-In PC1 Scenario 14 (o) -4 -2 0 2 In N-In PC1 Scenario 15 (p) *** -4 -2 0 2 In N-In PC1 Scenario 16 (q) Figure 3 Figure 4 −10 −5 0 5 −5 0 5 Ph LDMC An gs Ci iWUE Fv_Fm AbMF BgMF PC1 (35.78%) PC2 (21.8%) Figure S1 Figure S2 Figure S3 Figure S4