Introduction

The benefits of micropropagation can be valued with successful plant growth in the greenhouse (Hazarika 2003). The regulation of environmental conditions including light intensity, CO2, sucrose concentration, and nutrients supplied during in vitro culturing, may improve the plant growth of certain species when they are transferred to ex vitro conditions (Adelberg et al. 2015; Barrales-López et al. 2015). Since growth of plant organs has different nutrient requirements (Ramage and Williams 2002; El-Hawaz et al. 2015a), use of MS medium (Murashige and Skoog 1962) for ex vitro subsequent growth may not be suitable (Adelberg et al. 2007, 2010). In addition, different plant genotypes might require specific in vitro nutrition (Williams 1993; Leifert et al. 1995; Gahan 2008). Multi-factor experiments using response surface methods (RSM) show the optimal mineral concentrations for ex vitro growth were different than in vitro growth for turmeric (Adelberg et al. 2013; El-Hawaz et al. 2015b). The concentrations of \({\text{PO}}_{4}^{3 - }\) and KNO3 in pre-hardening in vitro media have a greater impact than applying fertilizer on transplants grown in a soilless mix during a 2 week acclimatization in a greenhouse, however the random error due to variations in irrigation, wind and light distribution on the mist bed resulted in low accuracy models (Adelberg et al. 2013). Since most of the nutrients required for subsequent growth during acclimatization came from the medium used during in vitro pre-hardening, diluting the medium may limit the subsequent ex vitro growth Rooting and plant quality for acclimatization may be improved by lowering \({\text{NH}}_{4}^{+}\) in MS medium (Economou 2013). In turmeric, \({\text{NH}}_{4}^{+}\) free medium resulted in the best greenhouse growth during acclimatization, when \({\text{NO}}_{3}^{ - }\), \({\text{PO}}_{4}^{3 - }\), Ca2+ and Mg2+ were present at fairly high concentrations compared with MS (Adelberg et al. 2013).

The MS medium has sixteen elements, along with organic compounds and plant growth regulators (Murashige and Skoog 1962). Choosing which elements to vary in an experiment should be a deliberate strategy. Niedz and Evens (2007) clustered mineral salts in four groups: macronutrients (NH4NO3 and KNO3), mesonutrients (KH2PO4, MgSO4, and CaCl2), micronutrients (CuSO4, KI, CoCl2, H3BO3, and Na2MoO4) and chelated iron, and then screened the factors in high and low concentrations relative to MS using RSM. Using this design, MS mineral salts were screened for improved in vitro shoot and leaf quality of Rubus idaeus L. (Poothong and Reed 2014), Corylus avellana L. (Hand et al. 2014), and Pyrus cultivars (Reed et al. 2013a, b) during stage II of the micropropagation process. Other RSM experiments sufficiently identified the interactive effects of mesonutrient combinations on in vitro responses of Pyrus cultivars (Wada et al. 2013, 2015) and Curcuma longa (Adelberg et al. 2013; El-Hawaz et al. 2015a). Testing salts as factors caused ion-confounding results and the specific concentration effect of each element could not be specifically quantified (Niedz and Evens 2006, 2008). By balancing the charge of ions, Niedz and Evens (2008) determined a synergistic interaction of \({\text{NH}}_{4}^{+}\) and K+ that affected the in vitro growth of Citrus sinensis (L.) Osbeck. Low \({\text{NO}}_{3}^{ - }\) and low \({\text{NH}}_{4}^{+}\):K+ increased the growth of Brugmansia × candida Pers. (Niedz et al. 2012) and equimolar of \({\text{NH}}_{4}^{+}\):K+ enhanced the quality of Gerbera hybrid (Niedz et al. 2014). In a similar experimental design, the subsequent greenhouse growth of C. longa was maximized when \({\text{NO}}_{3}^{ - }\) was the only source of N in the in vitro pre-hardening medium (Adelberg et al. 2013).

Doubling the standard MS concentration of in vitro \({\text{PO}}_{4}^{3 - }\) from 1.25 to 2.5 mM increased the growth of Asparagus officinalis L. shoot apex culture (Murashige et al. 1972). With C. longa, phosphorus was the first mineral to be depleted from the 2.5 mM \({\text{PO}}_{4}^{3 - }\) medium, Murashige et al. (1972), during the 21 days micropropagation period (Adelberg 2010). The in vitro growth of Musa spp. was enhanced with raising \({\text{PO}}_{4}^{3 - }\) from 1.25 to 2 mM in solid MS medium (Amiri 2008). Solanum tuberosum L. in vitro growth and shoot length increased with 2.5 and 6.25 mM \({\text{PO}}_{4}^{3 - }\) in MS medium (Sarkar et al. 2004). Increasing the concentration of \({\text{NO}}_{3}^{ - }\) beyond the standard MS concentration improved the in vitro growth of Eucalupyus (Ivanova and Van Staden 2009), Prunus domestica L. (Nowack et al. 2007), and Eleutherococcus koreanum Nakai. (Lee and Paek 2011).

Liquid culture has inherent advantages for studying mineral nutrition when compared to semi-solid (agar-type) medium (De Klerk and Brugge 2011) including the elimination of gradients within the medium and greater water availability to the plant. High plant density culture is desirable for commercial utility (Leifert et al. 1991). In liquid cultures, increased density exhausted \({\text{PO}}_{4}^{3 - }\), K+, and N in medium and reduced N and \({\text{PO}}_{4}^{3 - }\) concentrations in Delphinium (Lumsden et al. 1990), Hemerocallis, and Iris tissue (Leifert et al. 1991). Spent medium analysis was used to show increasing in vitro plant density decreased \({\text{PO}}_{4}^{3 - }\) and Mg2+ in Hemerocallis tissue to levels that were below standards of adequacy for subsequent nursery growth (Adelberg et al. 2010). By including plant density as a factor in the RSM, increased plant density was shown to increase the need for more minerals in order for C. longa to grow faster when acclimatized in the greenhouse (Adelberg et al. 2013; El-Hawaz et al. 2015b).

Turmeric, C. longa, is a sterile perennial rhizomatous plant, and an economically important species within the Zingiberaceae family. Turmeric powder has a market value as a condiment food color and flavoring agent (Madan 2007) and is used in folk medicines of Asia and Middle East (Afifi and Abu-Irmaileh 2000; Latif et al. 2014). Turmeric was chosen as a model to develop media for herbaceous monocots used in phytochemical production or perennial ornamental plant micropropagation (Adelberg 2010).

The current research aims were (1) to observe the effects of low \({\text{NH}}_{4}^{ + }\) medium containing various mineral concentrations on the subsequent growth after 21 days acclimatization for two genotypes of C. longa (2) to identify optimal combinations of in vitro minerals and plant density for ex vitro relative gain in fresh biomass, leaf area index, and the shoot length, and (3) to compare these results to what is observed from liquid MS medium.

Materials and methods

Plant material

Curcuma longa genotypes L35-1 and L22-5 rhizomes were sterilized and initiated in stage I as described by Adelberg and Cousins (2006). The genotypes were supplied by the University of Arizona Southwest Center for Natural Products Research and Commercialization (UA Herbarium #375,742, ARIZ). Turmeric buds (0.5–1 cm) were micropropagated as stage II for 10 years by repeated subculture in 33 mL of MS medium including 100 mg myo-inositol, 0.25 mg nicotinic acid, 0.25 mg pyridoxine hydrochloride, 0.05 mg thiamine HCl, 6 % w/v sucrose and 1 µM benzyladenine (BA) in a cylindrical glass jar (180 mL) with a Magenta B-cap closure (Magenta Crop, Chicago, IL, USA) on an orbital shaker (100 rpm) at 23 ± 2 °C under cool white fluorescent lights with the intensity of 25–30 µmol m2 s−1 photosynthetically active radiant (PAR) for 16 h per day. Then each five buds were subcultured in 40 mL of 2 × SH medium (Schenk and Hilderbrandt 1972) supplied with 5 % w/v sucrose and 3 µM BA for three cycles 35 days each for adapting plantlets to the high mineral concentrations in treatment media.

Response surface methods for in vitro propagation

Turmeric buds were transferred into 40 mL of low ammonium (5 mM \({\text{NH}}_{4}^{+}\)) MS medium with 5 % sucrose, 3 µM BA (with vitamins and inositol, as above) and with twenty-five combinations of \({\text{PO}}_{4}^{3 - }\) (6.25, 8.25, 10.25 mM), KNO3 (18, 59, 100 mM), Ca2+ (2, 6, 10 mM) in different densities (50, 150, 250 buds/L) as shown in Table 1. To maintain the charge balance in media mixtures, \({\text{SO}}_{4}^{2 - }\) ion was used as a counter ion and considered as a covariant factor (Adelberg et al. 2013; El-Hawaz et al. 2015a). Media pH range was measured but not controlled (5.26–5.98), since adjusting pH by NaOH would introduce Na+ as a cofounding counter ion. At the same time, the genotypes were also propagated at the three densities in 40 mL of liquid basal MS medium (1962) with 5 % w/v sucrose and 3 µM BA as a control. All media included vitamins and inositol (as described above). Trimmed buds were subcultured for four 35 days cycles for adapting plantlets to treatment conditions and were grown on orbital shakers with lighting and temperature, as described above

Table 1 The twenty-five treatments combinations of four factors, in vitro plant density (buds/L), \({\text{PO}}_{4}^{3 - }\) mM, KNO3 mM, and Ca2+ mM and \({\text{SO}}_{4}^{2 - }\) mM was a covariate
Full size table

Media analysis

At the end of the fourth cycle, the concentration of \({\text{PO}}_{4}^{3 - }\), K+, Ca2+, Mg2+, \({\text{SO}}_{4}^{2 - }\), Fe3+, Mn2+, B3+, Zn2+, and Cu2+ in the remaining media were determined by Thermo Jarrell Ash Model 61E Inductively Coupled Plasma (ICP; Analytical West, Inc., Corona, CA, USA). A Flow Injection Analyzer (FIAlab-2500; FIAlab instruments, Inc., Bellevue, WA, USA) determined N in \({\text{NO}}_{3}^{ - }\). The Agricultural Service Laboratory at Clemson University (Clemson, SC, USA) provided the mineral analysis.

Acclimatization

Following the fourth cycle, the rooted in vitro plantlets from each vessel were washed with tap water and assigned to planting trays (601, six sections L × W × H, 16 × 12 × 5.5 cm polyethylene insert, fitting a webbed 1020 polyethylene tray, Landmark Plastics, Akron, OH, USA) based on the initial plant density as follows: the products of low plant density (50 buds/L) were transferred to one section, the products of moderate plant density (150 buds/L) were transferred to 3 sections, and the products of high plant density (250 buds/L) were transferred to 5 sections. Plants grew for 21 days in soilless mix Fafard 3B (Canadian sphagnum peat moss, 7.62/20.32 cm processed pine bark, perlite, vermiculite, wetting agent, starter fertilizer and dolomitic limestone; Sun Gro Horticulture, Agawam, MA, USA) in the greenhouse (latitude = 34.67350, and longitude = −82.83261) with mist frequency of 8 s every 10 min during the daylight hours. The trays were moved one position in series around the mist bed, shifting daily to compensate for non-uniform irrigation, wind and shading on the greenhouse mist bench (Fig. 1).

Fig. 1
figure 1

Daily movements of trays distributed error due to light and misting irregularities in the greenhouse

Full size image

Greenhouse growth

After a 21 days acclimatization period, harvested plants were washed free of the soilless mix and dried on paper towels. The relative fresh biomass gain (RFBG) was calculated for the harvested plants/vessel as Fresh biomass after 21 days out/Fresh biomass at day 0 in the greenhouse.

The shoot length and leaf area were measured from flat images using Image J software (Image J/FIJI 1.46, Rasband, W.S., ImageJ, U.S. National Institutes of Health, Bethesda, Maryland, USA, http://imagej.nih.gov/ij/, 1997–2014). A leaf area index (LAI) was determined from each vessel by dividing the number of initial buds in a vessel into the sum total leaf area of all plants in the vessel (cm2/bud). The average shoot length for each vessel was determined for shoots above a 4 cm height.

Response surface methods design and statistical analysis

D-optimal criteria were used to select 25 treatments (out of 81 possible combinations of four factors at three levels each). This set of treatment combinations was duplicated to examine an extra qualitative factor (genotypes) at two levels. The factor terms to include in the response surface models (Tables 2, 3, 4) were selected by forward stepwise methods when P < 0.05. Statistical criteria including R 2, the adjusted R 2 (\(R_{a}^{2}\)), the predicted R 2 (\(R_{p}^{2}\)), the overall F statistic, and P value, were used to assess the utility of each model. The MS control experiment was run concurrently to test the effects of two genotypes and the three levels of plant densities in the basal MS medium (1962) in a completely randomized experiment (2 × 3 factorial arrangement with the three replicates). The factorial model terms were considered significant at P < 0.05. The experimental designs, analysis, and graphs were created using JMP version 11.0 (SAS Inst., Cary, NC, USA).

Table 2 Turmeric relative fresh biomass gain (RFBG) was affected by in vitro plant density, KNO3, and Ca2+ after 21 days acclimatization period
Full size table
Table 3 Turmeric leaf area index (LAI) produced after 21 days acclimatization period was affected by in vitro plant density, KNO3, Ca2+, and genotypes
Full size table
Table 4 Turmeric shoot length was affected by in vitro plant density, \({\text{PO}}_{4}^{3 - }\), KNO3, Ca2+, and genotypes after 21 days acclimatization period
Full size table

Results and discussion

Relative fresh biomass gain (RFBG)

During the 21 days acclimatization period, fresh biomass from plantlets grown on basal MS medium more than doubled (2.6 ± 0.5 fold for both genotypes), regardless of plant density. In treatment media, subsequent RFBG could be optimized by selecting KNO3 concentration in the tested range, as indicated by the significant negative quadratic term (Table 2). The interaction of KNO3 with plant density and with Ca2+ significantly affected the subsequent RFBG during the 21 days acclimatization period of both genotypes (Table 2). Plantlets grown at high plant density required 42 mM KNO3 and 2 mM Ca2+ in vitro to increase subsequent RFBG 3.3 ± 0.5 fold. Plantlets grown at low plant density and subjected to 18 mM KNO3 and 2 mM Ca2+ in vitro resulted in the greatest subsequent RFBG (3.7 ± 0.5 fold, Fig. 2). Choosing KNO3 as a macronutrient salt to charge balance \({\text{NO}}_{3}^{ - }\) made it difficult to distinguish between the effects of \({\text{NO}}_{3}^{ - }\) and K+. Although K+ and \({\text{NO}}_{3}^{ - }\) were supplied in equimolar amounts, spent medium analysis showed twice as much K+ remained following the culture period (Fig. 3). This suggests that regardless of \({\text{NO}}_{3}^{ - }\) concentration, the concentration of K+ was more than sufficient for maximum growth, and the increased demand for KNO3 in high-density culture was driven by the increased need for \({\text{NO}}_{3}^{ - }\).

Fig. 2
figure 2

Effect of in vitro plant density and KNO3 concentration on the relative fresh biomass gain of turmeric after 21 days acclimatization period. The counter surface and the surface plus residual points were presented when Ca2+ concentration was set at 2 mM. The line segment shows MS relative fresh biomass gain plotted to 40 mM \({\text{NO}}_{3}^{ - }\)

Full size image
Fig. 3
figure 3

A regression plot shows the concentration of K+ and N (mmol/vessel) in spent media after 35 days of turmeric micropropagation. The regression equation is K+ in remaining media = 0.062659 + 1.9076067 × N in remaining media was evaluated with R 2 = 0.958

Full size image

In a similar study, turmeric subsequent RFBG after a 14 days acclimatization period had been affected by mineral nutrition in vitro, but the non-uniformity in mist bed conditions, wind, light intensity, and irrigation, may have caused uncontrolled error that resulted in the model having low reliability and predictability (Adelberg et al. 2013). Water loss from the leaf surface during acclimatization severely limits growth, and controlling relative humidity, air movement and light intensity in the acclimatization environment is necessary to allow in vitro plantlets to grow well in a greenhouse (Kozai 1991). In the present study, the variation in light, wind, and irrigation water on the greenhouse bench was randomly distributed across treatments by rotating the trays in the greenhouse one position every day.

Leaf area index (LAI)

In vitro plant density significantly affected subsequent LAI of plantlets from MS after the 21 days acclimatization period. Both genotypes had 22 ± 8 cm2/bud from the high plant density, which was significantly lower than the subsequent LAI from the low plant density (36 ± 8 cm2/bud). In the treatment media, the significant negative quadratic term and the interaction of plant density with KNO3 indicated that the optimal concentration of KNO3 was in the range of the experiment and was affected by plant density (Table 3). After 21 days acclimatization, the treatment media level of Ca2+ significantly affected subsequent turmeric LAI (Table 3) in both genotypes. In vitro medium with 70.4 mM KNO3 and 2 mM Ca2+ and at high plant density resulted in subsequent turmeric LAI of 67 ± 16.4 cm2/bud. In comparison, treatment medium with 50.7 mM KNO3 regardless of Ca2+ and at low plant density, maximized subsequent turmeric LAI (78.4 ± 16.4 cm2/bud) (Fig. 4). In the optimal media, N concentration was still in a similar range to that found in MS medium except \({\text{NO}}_{3}^{ - }\) was higher in the optimal media since \({\text{NH}}_{4}^{ + }\) had been reduced from 20 mM to 5 mM, and K+ was used as a counter ion instead of \({\text{NH}}_{4}^{ + }\) or Na+. Leaf area and shoot length of field grown turmeric correlated well with final rhizome yield (Panja et al. 2002; Ravindran et al. 2007 and citation within).

Fig. 4
figure 4

Effect of in vitro plant density and KNO3 concentration on the leaf area index (cm2/bud) of turmeric after 21 days acclimatization period. The counter surface and the surface plus residual points were presented when Ca2+ concentration was set at 2 mM in vitro. The line segment shows the leaf area index of plants from MS at high and low plant density plotted to 40 mM \({\text{NO}}_{3}^{ - }\)

Full size image

Shoot length

Genotypes grown on MS medium had different shoot lengths. Genotype L22-5 elongated to 6.0 ± 0.6 cm regardless of plant density after 21 days acclimatization. Shoot length of genotype L35-1 elongated to 4.0 ± 0.6 cm from high plant density and to 6.0 ± 0.6 cm from low plant density. Similar to RFBG and LAI from the treatment media, the subsequent turmeric shoot length could be optimized by selecting KNO3 in the tested range as indicated by the significant negative quadratic term (Table 4). Plant density and Ca2+ significantly affected the shoot length regardless of the genotype (Table 4). The RSM identified that high plant density growth with medium containing 64 mM KNO3, 6.25 mM \({\text{PO}}_{4}^{3 - }\), and 2 mM Ca2+ resulted in shoot length of 8.0 ± 0.9 cm after a 21 days acclimatization period, which was shorter than plants from low plant density in the same medium (10.9 ± 0.9 cm, Fig. 5). The negative quadratic term of KNO3 indicates that the optimal concentration is in the range that was tested. In the present study, the optimal in vitro treatment concentration of K+ was 3–4× greater and \({\text{NO}}_{3}^{ - }\) was 1.5× greater than concentrations specified in MS.

Fig. 5
figure 5

Effect of in vitro plant density and KNO3 concentration on the shoot length (cm) of turmeric after 21 days acclimatization period. The counter surface and the surface plus residual points were presented when other in vitro factors were set at 6.25 mM \({\text{PO}}_{4}^{3 - }\) and 2 mM Ca2+. The line segment shows shoots length from MS plants in low and high plant density plotted to 40 mM \({\text{NO}}_{3}^{ - }\)

Full size image

With proper combination of factors, plants from treatment media were larger 21 days after acclimatization than those grown in MS (Fig. 6). The differences in RFBG, LAI, and shoot length were plainly observed when viewing the young plants. Many scientists have used the half MS medium for the third stage of micropropagation (shoot development) to avoid the high N concentration in the medium (especially \({\text{NH}}_{4}^{ + }\)) that may cause hyperhydricity (Ziv 1991; George et al. 2008; Ivanova and Van Staden 2009; Reed et al. 2013b) and to improve in vitro rooting of transferring plants (Moncousin 1991; Economou 2013). However, diluting medium concentration could result in inadequate concentrations of other biologically significant minerals (Adelberg et al. 2013).

Fig. 6
figure 6

Two-dimensional images of Curcuma longa after 21 days of ex vitro growth to show the effect of in vitro growth media. Low plant density on MS medium a genotype L35-1 and b genotype L22-5 compared to plants grown at low plant density, 10.25 mM \({\text{PO}}_{4}^{3 - }\), 2 mM Ca2+, and 59 mM KNO3, c genotype L35-1 and d genotype L22-5. Plants grown in vitro on MS medium at high plant density, e genotype 35-1 and f genotype L22-5 compared to the treatment of high plant density, 10.25 mM \({\text{PO}}_{4}^{3 - }\), 10 mM Ca2+, and 100 mM KNO3, g genotype L35-1 and h genotype L22-5. As a consequence of using D-optimal selection, less than one-third (25/81) of the treatment points were tested thus treatment points photographed here were closest to the treatment highlighted

Full size image

The concentration of Ca2+ in the optimal media (2 mM) was lower than the concentration specified in MS (3 mM). In the cytosol, free Ca2+ is a signal transducer for many enzymes. Thus Ca2+ concentration is highly regulated and is transported out of the cytosol by energy-dependent efflux (Taiz and Zeiger 2006). In the treatment media, the uptake of Ca2+ and plant tissue Ca2+ concentrations increased with increasing Ca2+ concentration (data not shown). During acclimatization, the photosynthetic process might not provide adequate energy to maintain efflux of Ca2+ from the cytosol. Reducing the strength of MS medium to half might enhance the subsequent acclimatization growth of in vitro plants with lowered Ca2+and less \({\text{NH}}_{4}^{ + }\), but more KNO3 and \({\text{PO}}_{4}^{3 - }\) are needed for stage III pre-hardening.

Conclusion

In the present study, KNO3 was the most important nutrient factor that increased subsequent ex vitro growth from high plant density culture. The residual K+ in the remaining media indicated that the supply of K+ in the higher concentrations might not be necessary as a nutrient and was tolerated by rapidly growing plants at high density. Potassium cation was useful for increasing \({\text{NO}}_{3}^{ - }\) in a salt-based formulation with reduced \({\text{NH}}_{4}^{ + }\). Murashige and Skoog (1962) relied on NH4NO3 to provide one-third of the \({\text{NO}}_{3}^{ - }\) concentration.

A low \({\text{NH}}_{4}^{ + }\) (5 mM) stage III medium that was developed in prior works to promote turmeric growth during acclimatization that also had a high \({\text{PO}}_{4}^{3 - }\) concentration (Adelberg et al. 2013; El-Hawaz et al. 2015a). In this study, the \({\text{PO}}_{4}^{3 - }\) concentration tested were 5× to 10× greater than the 1.25 mM \({\text{PO}}_{4}^{3 - }\) used in basal MS medium. The fact that \({\text{PO}}_{4}^{3 - }\) did not influence ex vitro quality metrics, suggests that 6.25 mM \({\text{PO}}_{4}^{3 - }\) is adequate for turmeric during subsequent acclimatization. Turmeric grew larger and faster during the ex vitro acclimatization period when subjected to high \({\text{PO}}_{4}^{3 - }\), high KNO3, low Ca2+ and low \({\text{NH}}_{4}^{ + }\) media in comparison to the MS medium. The optimal concentrations depended on the density of plants in culture, with greater densities preferred for commercial application and lesser density yielding the largest plants.

In vitro mineral nutrition is a critical factor for the subsequent acclimatization stage before the root system is established in the soilless mix. Selecting an optimal media requires defining a high or low density application, and choosing which ex vitro responses are most important. Only after these decisions are made, can a nutrient medium be formulated from inorganic salts.

Using RSM allows for modification of multiple factors within in vitro media for better ex vitro growth. In a plant propagation nursery, once the plant has attained the proper size, it is removed from the subsequent mist and repotted to a larger container. The rapid removal of acclimatized plants under mist reduces the risk of pest infestation and allows an early application of conventional fertilizer, therefore increases the efficiency of the propagation nursery. In vitro treatments that make ex vitro plants grow more quickly through acclimatization are superior as propagation stock.