Nepenthes pitchers are CO2-enriched cavities, emit CO2 to attract preys

Carnivorous plants of the genus Nepenthes supplement their nutrient deficiency by capturing arthropods or by mutualistic interactions, through their leaf-evolved biological traps (pitchers). Though there are numerous studies on these traps, mostly on their prey capture mechanisms, the gas composition inside them remains unknown. Here we show that, Nepenthes unopened pitchers are CO2-enriched ‘cavities’, when open they emit CO2, and the CO2 gradient around open pitchers acts as a cue attracting preys towards them. CO2 contents in near mature, unopened Nepenthes pitchers were in the range 2500–5000 ppm. Gas collected from inside open N. khasiana pitchers showed CO2 at 476.75 ± 59.83 ppm. CO2-enriched air-streaming through N. khasiana pitchers (at 619.83 ± 4.53 ppm) attracted (captured) substantially higher number of aerial preys compared to air-streamed pitchers (CO2 at 412.76 ± 4.51 ppm). High levels of CO2 dissolved in acidic Nepenthes pitcher fluids were also detected. We demonstrate respiration as the source of elevated CO2 within Nepenthes pitchers. Most unique features of Nepenthes pitchers, viz., high growth rate, enhanced carbohydrate levels, declined protein levels, low photosynthetic capacity, high respiration rate and evolved stomata, are influenced by the CO2-enriched environment within them.

Scientific RepoRts | 7: 11281 | DOI:10.1038/s41598-017-11414-7 the atmosphere are 159 and 0.30 mm Hg, respectively. We also detected CO 2 dissolved in N. khasiana pitcher fluid by headspace GC-MS (VF-5 column, ret. time 1.65 min; EI-MS, m/z: 44 (M + ), 32). Mass data of CO 2 from the pitcher fluid matched with its authentic standard. We measured the pH of unopened N. khasiana pitcher fluid as 3.54 ± 0.09 (n = 4), and on prey capture the fluid became more acidic (pH 2.47 ± 0.25, n = 4). CO 2 , stomata in Nepenthes pitchers. In SEM images, we found N. khasiana leaves (laminae) hypostomatic i.e., stomata observed only in their abaxial (lower) sides (Fig. 3a), and not in adaxial (upper) sides. But N. khasiana pitchers (both unopened and open pitchers) showed stomata in their outer sides, and 'modified stomata' in their inner sides (Fig. 3b-e). No stomata were seen at the inner sides of N. khasiana pitcher lids (Figs S8-S10). Stomata in the abaxial sides of the leaves and at the outer sides of pitchers were normal ones with two guard cells ( Fig. 3a-c), whereas stomata inside the pitchers were modified 'lunate cells' , pointing downwards, with only one guard cell (Figs 3d,e and S3). These modified stomata inside the pitcher were found embedded in crystalline epicuticular wax layers (Fig. 3d,e). CO 2 , trichomes, prey capture. Leaf abaxial and adaxial sides of N. khasiana showed only glandular trichomes (data not shown) at a low density. Branched non-glandular and glandular trichomes were observed on N. khasiana tendrils (partially seen in Fig. 3h), at the outer sides of their pitchers and upper sides of their lids (Figs 3b and S11-S13), glandular trichomes only were found in the inner sides of lids (Figs S8-S10), and no trichomes were observed in other inner sides of pitchers (peristome, slippery and digestive zones) (Figs 3d-g and S2-S4).

Discussion
Our data demonstrate that Nepenthes unopened pitchers are CO 2 -enriched 'cavities' , when lids open they release CO 2 at a high 3000-5000 ppm to an ambient ~400 ppm atmosphere, and then continue releasing CO 2 resulting in its gradient surrounding them. Weight (difference) measurements of Nepenthes pitchers indicate the release of a denser gas (CO 2 ; density CO 2 /air 1.980/1.225 kg/m 3 ) within them, filled at a slightly higher pressure compared to the atmosphere. Nepenthes pitchers generally stay in 'upright' position, and the gas within is emitted through the pitcher-lid opening. Open pitchers of N. khasiana are constant emitters of CO 2 (476.75 ± 59.83 ppm, n = 6), a sensory cue. Most insects pay special attention to 'subtle variations' or 'gradients' of CO 2 in the form of plumes arising from individual point sources 14,15 . Insects have well developed CO 2 receptors which can detect these variations (even small variations) as a means of locating their food 14 . Moreover, CO 2 emitting devices are widely used as traps against mosquitoes, flies and other insects 16 . In this study, on CO 2 -streaming (1% CO 2 -enriched air) for 12 days through N. khasiana pitcher tops, we found a substantial increase in aerial preys (insects) captured within them (preys captured = 31.17 ± 11.91, n = 6), compared to air-streamed control pitchers (same flow rate; preys captured = 16.20 ± 5.15, n = 6) and unmodified (normal) pitchers (preys captured = 19.67 ± 5.43, n = 6) (Fig. 1). These counts are excluding the ants (dead) crawled into these pitchers from the ground, and the ants count did not show any pattern between the CO 2 -enriched, air-streamed and unmodified (normal) pitchers. In Fig. 1, the insect capture rates in these three different experimental conditions (CO 2 -enriched, air-streamed and unmodified pitchers) are proportional to the CO 2 emission rates from N. khasiana pitcher tops. These data demonstrate CO 2 as an insect attractant emitted by Nepenthes prey traps, and reveals a new prey capture mechanism within them.
Most Nepenthes species secrete pitcher fluids with viscoelastic properties. Fluids in unopened pitchers are sterile 17 , and once open microbes and inquilines invade them. Our results show that a high level of CO 2 is dissolved in N. khasiana pitcher fluids (Fig. 2). Open, prey captured pitcher fluids showed low levels of O 2 (Fig. 2), and very low (even anoxia) or decreasing levels of oxygen were reported in Sarracenia purpurea, Utricularia and Genlisea traps [18][19][20] . Dissolved CO 2 in Nepenthes pitcher fluid instantaneously forms equilibrium with its hydrated form H 2 CO 3 which dissociates into H + and HCO 3 − 21 . The relative changes to any one of these molecules/ions control the pH and optimum activity of the digestive enzymes secreted into the pitcher fluid by specialized glands 21 (Fig. 3f,g). N. khasiana pitcher fluids are acidic, and lid opening and prey capture further reduce the pH. Similar pH trends were observed in pitcher fluids of several Nepenthes species 4,5,11,22 . pH reduction on prey capture (even after release of the high levels of CO 2 on pitcher opening) is critical for optimum enzyme activity (prey digestion) and absorption of nutrients, and this is achieved through a proton (H + ) pump [22][23][24] . This acidic pH could also be controlling the growth of pitcher inhabitants (microbes, mosquito larvae, small aquatic organisms etc.). CO 2 dissolved in the pitcher fluid is one of the factors making it acidic and it also acts as a preservative to the pitcher fluid.
Once Nepenthes pitchers become mature, their 'tight lid sealings' 1 open and release the elevated CO 2 within, making them ready for prey capture. The sequential events of lid opening, CO 2 release and prey capture are sensed by these plants, and they release antifungal naphthoquinones (droserone, 5-O-methyl droserone, plumbagin, 7-methyl juglone) into the pitcher fluid (Figs S5-S7), preventing infections from incoming preys 5 .
Stomata are small pores controlling gas exchange, mainly CO 2 and water vapour, found in leaves and other organs in plants 25 . Stomata inside N. khasiana pitchers were 'modified' , pointing downwards, with only one guard cell (Figs 3d,e and S3). Similar modified stomata embedded in wax crystals were observed by SEM studies in the inner sides of pitchers of N. rafflesiana 11,12,22 , N. alata 1,3,26 , N. mirabilis 3 , N. diatas 23 and other Nepenthes species/ hybrids 27,28 . Most authors described these stomata as 'transformed' or as 'lunate cells' with a convex structure in the inner surface of Nepenthes pitchers, and explained this modification as an evolutionary adaptation contributing to prey capture by disrupting the adhesion of insect feet and blocking entry of their claws 12,22,23,26 . Owen and Lennon, 1999 suggested the function of this 'modified stomatal complex' as 'water secretion' or 'gas exchange' or even as a 'mystery' 1 . But, absence of 'pores' 1,22 in these 'modified stomatal structures' nullifies the chances of them functioning as vents in 'gas exchange' .
Similar to our observations, Pavlovič and co-workers reported stomata on the abaxial sides of laminae of N. alata and N. mirabilis, and very low stomatal density in Nepenthes pitchers 3 . Other studies also reported modified stomata at the interior of pitchers and overall low stomatal density in pitchers of various Nepenthes species 1 . Stomatal distribution in laminae (abaxial, high) and pitchers (stomata with two guard cells, outer side; low) are  Table S2). matching with their photosynthetic capacities, high (laminae) and very low (pitchers). In most cases, we found high density of the 'modified stomata' at the pitcher inner (top) sides (Fig. S3) 1,9,23 . In Nepenthes, pitchers are formed by the folding of leaves with their adaxial (upper) surfaces turning into inner sides of these traps 1 . It is significant that, the leaf upper surfaces are devoid of stomata, but the pitcher inner surfaces 'evolved' these 'modified stomata' (Fig. 3d,e). In pitchers of Sarracenia, Darlintonia, Heliamphora and Cephalotus, stomata (normal) are found in their outer surfaces or in their lids/'hoods' , and 'stomata-like structures' present within their pitcher tubes are 'permanently open' and not 'functional' 12 . It is proven that increase in CO 2 even in the range of 100 ppm has a profound effect on the stomata (modifies their morphology) in plants 14 . The transformed stomatal aperture with a single guard cell (Fig. 3d,e) at the interior (only) of Nepenthes pitchers is most probably a manifestation of the high CO 2 (approx. 4000 ppm, nearly 10 times the ambient) atmosphere within them. But, evidences gathered so far are not conclusive on the function of these 'modified stomata' or 'lunate cells' (Figs 3d,e and S3) 1,12 . Crystalline epicuticular wax in thick layers, as observed in the upper part of inner pitcher walls of N. khasiana and several other Nepenthes species, is not distinctly seen in other portions of the pitchers (lid, peristome, liquid zone, outer surface) and in the abaxial and adaxial sides of their leaves (Fig. 3). These inner waxy layers define the hydrophobic slippery zone, which minimizes insect attachment. Recent evidences also demonstrate high level of CO 2 as a factor which enhances cuticular wax density in plants 29 . Nepenthes prey traps display a unique natural model of evolution of stomata in a CO 2 -enriched atmosphere.
Trichomes, a group of epidermal microstructures, carry out diverse functions in plants, and in carnivorous plants one of their roles is facilitating prey capture 30,31 . In fact, relatively high density of branched trichomes was observed at the top outer sides N. khasiana pitchers and their lids 1 (Figs S11-S12), and no trichomes were observed in deep interior of the pitchers. But, significantly, Sarracenia, Heliamphora, Darlingtonia and Cephalotus pitchers have trichomes in their interior zones, including their innermost digestive zones 1,12,32 . Branched trichomes on the exterior of Nepenthes pitchers (and their lids) provide a foothold to the visitors (termites, ants etc.) 30,31 , enhancing the chances of their ultimate 'luring' to the interior of the traps. Edible trichomes in N. albomarginata are known to 'lure' termites into their pitcher traps 23,33,34 . Elevated CO 2 within Nepenthes traps could be one factor reducing the trichome density (particularly branched ones) in the inner sides of Nepenthes pitchers 35 .
SEM micrographs showed numerous vascular bundles within the roots and tendrils of N. khasiana (Fig. 3h-l), but no gas flow was detected from tendril (cross-section) into the pitcher cavities. Respiration (dark) rates of non-carnivorous herbaceous plants are typically less than 50% of their photosynthetic rates, but, the average respiration/photosynthetic rate in terrestrial carnivorous plants is as high as 63% 36 . Again, the traps (pitchers, snap trap) of terrestrial carnivorous plants (Nepenthes, Sarracenia, Dionaea muscipula) showed much higher respiratory costs (respiration/photosynthetic rate 158%) than their laminae (lamina, phyllodia, petiole) (respiration/photosynthetic rate 19%) 36 . More evidences for higher respiration rates (in traps compared to laminae) are available in carnivorous plants with 'active' trapping mechanisms (D. muscipula; Utricularia, bladder traps) [36][37][38] . Our results show that, N. khasiana laminae have significantly higher photosynthetic capacity compared to their pitchers whereas respiration rates are comparatively high in pitchers. Similarly, maximum quantum yield of PSII (Fv/Fm) in N. khasiana laminae is high compared to their pitchers. These parameters are matching with similar previous measurements in other Nepenthes species 39 . Unlike most plant leaf structures, high growth rate and unique physiological functions (prey attraction, capture, digestion, absorption of nutrients) of Nepenthes pitchers demand more energy, prompting higher respiration rates in the trap tissues, resulting in the release of more of CO 2 . Carnivorous plants follow the C3 photosynthetic pathway, and high CO 2 levels are also known to enhance respiration rates in C3 plants 40 . Thus, we demonstrate respiration of pitcher tissues as the factor contributing to the high CO 2 within the 'closed cavities' of Nepenthes traps.
Nepenthes tendrils and pitchers grow at a faster rate from their leaf terminals. 'Rapid elongation' of growing Nepenthes pitchers and their limited growth after opening of the lid sealing were previously observed by other authors 1 . Owen and Lennon, 1999 found a uniform growth rate of 0.0147 ± 0.0001 cm per h (0.35 cm per day) for N. alata pitchers, from initiation to the point of lid opening 1 . A small incision on defined N. khasiana pitchers (initial length, 6-8 cm) released the high CO 2 within them, and these pitchers continued growth at a diminished rate compared to control pitchers (Fig. 4). In control (uncut) pitchers, the balancing of CO 2 levels (with atmosphere) occurs only on lid opening. Our data indicate that, as in other CO 2 -enrichment studies, elevated (entrapped) CO 2 within acts as a growth promoter of Nepenthes prey traps. Recent studies revealed key data/ facts on comparative anatomy 41 and construction costs 42 of leaves/pitchers of Nepenthes species, leaf development in S. purpurea 43 and the influence of CO 2 on leaf phenology in plants 44 . More investigations, in the light of the discovery of CO 2 within, could possibly unravel similar growth patterns (tissue specific changes in cell division) 43 and faster growth rates in Nepenthes pitchers. Carbon contents of N. khasiana leaves are comparable to those of non-carnivorous plants 3,42 , but, both C and N contents are comparatively low in the pitchers 42 . As in other Nepenthes species 3, 45, 46 , the C/N ratio of N. khasiana pitchers is high, 26.05 (n = 4). CO 2 (high) and CO, CH 4 and N 2 O (ambient) found in Nepenthes pitchers are greenhouse gases. Global CO 2 levels are predicted to go up to 800 ppm by 2100 and further onto even higher levels 14 . Nepenthes prey traps with elevated CO 2 contents (3000-5000 ppm) are simulating this futuristic scenario in their 'closed cavities' (before trap opening). As in other CO 2 -enrichment experiments 14 , high carbohydrate and low protein contents were detected in Nepenthes pitchers 3 . Carbohydrate accumulation is a major acclimation response to elevated CO 2 14 . High carbohydrate contents in pitchers, transformed into nectar by nectaries (Figs S3 and S4), act as a major 'lure' in prey capture. Chlorophyll content is generally low in pitchers compared to their laminae. In some Nepenthes species, pitchers are red-tinted indicating low chlorophyll contents (Fig. S1). Pitchers in Nepenthes have very low photosynthetic rates compared to their laminae 3 . Reduction in photosynthetic rates in Nepenthes pitchers is primarily due to factors such as replacement of chlorophyll-containing cells with digestive glands, low nitrogen, chlorophyll contents and low stomatal density 3,14 . Photosynthetic Nitrogen Use Efficiency (PNUE) is also significantly low in Nepenthes pitchers compared to their laminae. Recently Pavlovič and Saganová pointed out reduced Rubisco activity in Nepenthes prey traps 39 , and Rubisco content is known to decrease with elevated CO 2 . These factors viz., photosynthetic rate, C/N ratio, carbohydrate/protein contents, chlorophyll content and PNUE, of several Nepenthes species were compared between their laminae and pitchers by various groups (N. alata and N. mirabilis 3 , N. talangensis 47 46 ). These parameters of Nepenthes leaves and pitchers were also compared to non-carnivorous plants 45,46 .
These trends in Nepenthes pitchers mainly, burst of growth, enhanced carbohydrate levels, declined protein levels, drop in photosynthetic capacity, high respiration rate and evolved stomata, are probable manifestations of the enhanced CO 2 atmosphere within them. These evidences also infer Nepenthes pitchers as ideal examples reflecting the effects of an anticipated high CO 2 level on Earth's surface, on the characteristic features of plants.
Recently, several groups put forward 'construction cost or cost/benefit theories' 3,42,45,46 on Nepenthes prey traps. Most of these studies estimated the nutritional benefit gained from captured preys above (at least marginally) the cost of constructing traps by leaf modification. Future construction cost estimates need to take into account of the acclimation responses of Nepenthes pitchers due to the 'so far unknown factor' of high CO 2 content within them.
In conclusion, Nepenthes pitchers are CO 2 -enriched cavities, and CO 2 emission from open pitchers acts as a sensory cue attracting insects towards these traps. Most of the characteristic features of Nepenthes pitchers are influenced by the high content of CO 2 entrapped within them. This study also hypothesizes Nepenthes pitchers as natural model systems mimicking an anticipated elevated CO 2 scenario on Earth.
Lids of mature (about to open, red colour appears at the peristome portion) N. khasiana pitchers in the field were sealed with super glue (to prevent lid opening). Then a small 'cut' (average 5.4 × 5.7 mm) was made on the top half (above liquid zone) of the pitcher (for gas release). After 24 h, the cut portion was sealed with parafilm/ super glue. After 2 days of sealing, pitchers were collected and subjected to gas analysis. In another set of experiments, lids of opened N. khasiana pitchers (opened a day before) were sealed back with super glue. After 2 days of sealing, these pitchers were collected and their gas compositions were analyzed.
Gas analysis by GC-FID/ECD/TCD. N. khasiana/Nepenthes hybrid unopened pitchers were opened underwater and the gases inside pitchers were collected by the displacement of water. This is to avoid possible mixing with air and dilution of the contents of the pitchers, when opened in air. The gases from the pitchers were transferred to syringes and analyzed through gas chromatography. A Clarus 580 gas chromatograph (Perkin Elmer, Waltham, USA) equipped with a Flame Ionization Detector (FID) and an Electron Capture Detector (ECD) was used. FID had a Methanator for converting CO and CO 2 to methane. ECD measured nitrous oxide in the sample. A gas sampling valve with 100 µl sampling loop was used for injecting the sample to the column. Isothermal separation was achieved at 35 °C in an Elite-PLOT Q column (30 m × 0.53 mm) with nitrogen carrier gas. Another NUCON 5765 gas chromatograph (Aimil, New Delhi, India) with a Thermal Conductivity Detector (TCD) and packed column (PORAPAK Q, 80/100 mesh, 5 m long) with nitrogen as carrier gas was used for the measurement of oxygen in the samples. FID, Methanator and ECD were calibrated with the standard gas mixture containing CH 4 , CO 2 , CO and N 2 O in nitrogen gas. SEM of N. khasiana roots, leaves, tendrils and pitchers. SEM analyses of N. khasiana abaxial/adaxial sides of leaves, inner/outer sides of pitchers, lids, tendril and roots were carried out on a S-2400 Scanning Electron Microscope (Hitachi, Tokyo, Japan) (Figs 3, S2-S4, S8-S13). N. khasiana samples were fixed with 3% gluteraldehyde in phosphate buffer and kept overnight. Samples were then dehydrated sequentially with 30%, 50%, 70% ethanol (15 min each, two changes) and 90%, 100% ethanol (30 min each, two changes). These dehydrated samples were subjected to critical point drying, coated with gold and viewed on the SEM. DART-MS of N. khasiana pitcher fluids. Pitcher fluids (yellow coloured) from prey captured N. khasiana pitchers (Fig. S5), chitin induced 5 (Fig. S6) and uninduced (colourless on opening, before prey capture) pitchers (Fig. S7) were collected, lyophilized and analyzed on an AccuTOF JMS-T100LC Mass Spectrometer having a DART (JEOL, MA, USA). Samples were analyzed directly in front of the DART source. Dry He was used at a flow rate of 4 L min −1 for ionization at 350 °C. Orifice 1 was set at 28 V, spectra were collected, and the data from 6-8 scans were averaged.
Nepenthes pitcher weight measurements. N. khasiana and Nepenthes hybrid (mature, unopened) pitchers were collected and their fresh weights were recorded. Then, pitchers were cut open just above the pitcher fluid level (to release the entrapped gas) and the entire pitcher contents were (very) quickly re-weighed (Table S1).
Field studies. N. khasiana pitchers were covered (netted) with colourless nets to prevent ants and insects entering on lid opening. Netting was done a week before opening on near mature pitchers. Three days after opening pitcher fluids were collected, lyophilized and analyzed. CO 2 -enriched air (1% CO 2 in air; Bhuruka Gases Ltd., Bangalore, India) was passed into just opened N. khasiana pitchers in the field through a small cut made above the fluid level by inserting a long, colourless tubing (inner diameter 2 mm; average flow 25.72 mL/min), and prey (aerial) capture was monitored for 12 days. Similarly, air at the same flow rate was streamed through control pitchers. On the 6th day, gas samples from inside test/control pitchers (just below the peristomes) were collected in syringes and analyzed by gas chromatography (n = 6, each). On the 12th day after lid opening, the entire contents of test/control N. khasiana pitchers (n = 6, each) were (separately) transferred to petri dishes (Fig. 1), and captured aerial preys (in each dish) were carefully counted. Similarly, prey (aerial) capture rates in normal (unmodified) pitchers (with no CO 2 /air streaming) in 12 days were also counted. In all three experiments, ants (dead) crawled into these pitchers from the ground were not considered (counted).
Tendrils of live N. khasiana plants were cut just below the pitchers and their cross sections were inserted into inverted syringes partially filled with water (for 6 days) in the field. On repeated experiments, no gas bubbling or any other changes in the water were observed. N. khasiana pitcher growth measurements. N. khasiana pitchers from the three populations in JNTBGRI garden sites with an initial growth of 6 to 8 cm were marked, their initial pitcher lengths were noted and small cuts (average 5.4 × 5.7 mm, to release the gas inside pitchers) were made above the fluid level. These test pitchers were constantly monitored, pitcher lengths on the day of lid opening and the number of days required till lid opening (from an initial stage of 6 to 8 cm) were noted. Similar measurements were also made on control N. khasiana pitchers (with no cuts) ( Fig. 4 and Table S2). Chlorophyll-A fluorescence, photosynthesis (A n ) and dark respiration (R d ) of N. khasiana laminae and pitchers. Chlorophyll-a fluorescence kinetics, A N and R D of N. khasiana laminae and pitchers were measured using a LI-COR 6400 XT portable infrared analyzer (LI-COR, Lincoln, NE, USA), equipped with a leaf chamber fluorometer. Laminae and pitchers from four N. khasiana plants in the field were subjected to these measurements. Fully grown N. khasiana laminae and healthy, prey captured pitchers (pitcher walls and lids were directly placed into the cuvette, independently) were taken for measurements. A constant PAR (photosynthetically active radiation) of 800 μmol m −2 s −1 of red (90%) and blue (10%) light was chosen as actinic light intensity and the measurement of chlorophyll fluorescence and P N were at ambient CO 2 level, temperature (33 ± 1 °C), RH (relative air humidity) ~80% and air flow rate of 300 μmol s −1 . R D was measured under similar conditions, except that the plant samples were under dark conditions. The laminae and traps were kept in the chamber for 5-10 min, until steady state of CO 2 concentrations were reached. Vapor pressure deficit in the sample cell ranged between 0.7 and 1.3 kPa. Minimal fluorescence (F 0 ) was measured for overnight dark adapted plant samples whereas maximal fluorescence (F m ) was recorded at a PAR of 8000 μmol m −2 s −1 (saturating flash). Maximal quantum yield of PSII was calculated as F v /F m = (F m − F 0 )/F m . Statistical analysis. Prey capture rates (Fig. 1), partial pressure measurements (Fig. 2), pitcher size/weight measurements (Table S1) and growth parameters of cut/uncut pitchers ( Fig. 4 and Table S2) are expressed as mean ± s.d. Statistical comparisons were done using student's t-test (Figs 1 and 2). Values of p < 0.05 were considered as statistically significant.

C, N contents in
Data availability. All data generated or analyzed during this study are included in this published article (and its Supplementary Information file).