• Open Access

Plastid targeting strategies for cyanophycin synthetase to achieve high-level polymer accumulation in Nicotiana tabacum

Authors


* Correspondence (fax 0049 3814983082; e-mail inge.broer@uni-rostock.de)

Summary

The production of biodegradable polymers in transgenic plants is an important challenge in plant biotechnology; nevertheless, it is often accompanied by reduced plant fitness. In order to decrease the phenotypic abnormalities caused by cytosolic production of the biodegradable polymer cyanophycin, and to increase polymer accumulation, four translocation pathway signal sequences for import into chloroplasts were individually fused to the coding region of the cyanophycin synthetase gene (cphATe) of Thermosynechococcus elongatus BP-1, resulting in the constructs pRieske-cphATe, pCP24-cphATe, pFNR-cphATe and pPsbY-cphATe. These constructs were expressed in Nicotiana tabacum var. Petit Havana SRI under the control of the constitutive cauliflower mosaic virus (CaMV) 35S promoter. Three of the four constructs led to polymer production. However, only the construct pPsbY-cphATe led to cyanophycin accumulation exclusively in chloroplasts. In plants transformed with the pCP24-cphATe and pFNR-cphATe constructs, water-soluble and water-insoluble forms of cyanophycin were only located in the cytoplasm, which resulted in phenotypic changes similar to those observed in plants transformed with constructs lacking a targeting sequence. The plants transformed with pPsbY-cphATe produced predominantly the water-insoluble form of cyanophycin. The polymer accumulated to up to 1.7% of dry matter in primary (T0) transformants. Specific T2 plants produced 6.8% of dry weight as cyanophycin, which is more than five-fold higher than the previously published value. Although all lines tested were fertile, the progeny of the highest cyanophycin-producing line showed reduced seed production compared with control plants.

Introduction

Transgenic plants may be an economical means of producing novel industrial feedstock, such as biodegradable polymers, if limitations to product accumulation and negative plant phenotypes can be overcome. The first such polymer produced in transgenic plants was polyhydroxybutyrate (PHB) (Poirier et al., 1992), a thermoplastic with properties similar to polyethylene and polypropylene. The genes for PHB synthesis were isolated from Ralstonia eutropha (Peoples and Sinskey, 1989a,b) and transferred to Arabidopsis thaliana (Nawrath et al., 1994; Bohmert et al., 2000), Brassica napus (Houmiel et al., 1999), Zea mays (Poirier and Gruys, 2002), Nicotiana tabacum (Lössl et al., 2003), Beta vulgaris (Menzel et al., 2003), Linum usitatissimum (Wrobel et al., 2004), alfalfa (Saruul et al., 2002) and Saccharum sp. (Petrasovits et al., 2007; Purnell et al., 2007). However, in high-PHB-producing plants, several negative effects on growth and development have been observed (Bohmert et al., 2000, 2002). The highest PHB accumulation was achieved in plants targeting the biosynthetic enzymes to the plastids of A. thaliana (Nawrath et al., 1994; Bohmert et al., 2000). When the genes required for PHB synthesis were introduced into the plastid genome of tobacco, the accumulation of PHB was associated with severe growth reduction and male sterility (Lössl et al., 2003).

Other polymers produced in transgenic plants include polyaminoacids, such as poly-γ-glutamate, poly-α-aspartate and poly-ɛ-lysine, which have a wide range of applications, e.g. as dispersants, thickeners or additives to hydrogels (Chang and Swift, 1999; Oppermann-Sanio et al., 1999; Oppermann-Sanio and Steinbüchel, 2002; Lössl et al., 2003). Polyaspartate is a soluble, non-toxic and biodegradable polycarboxylate (Tabata et al., 2000) that can be used to replace non-biodegradable polyacrylates in many industrial, agricultural and medical applications (Schwamborn, 1998; Joentgen et al., 2001; Zotz et al., 2001; Oppermann-Sanio and Steinbüchel, 2002). Because no polyaspartate-producing organism has been identified to date, the polymer is chemically synthesized (Schwamborn, 1996). However, it can also be obtained from cyanophycin (multi-l-arginyl-poly-l-aspartic acid). Cyanophycin is a cyanobacterial reserve polymer composed of a poly-α-aspartic acid backbone with arginine residues linked via their α-amino group to the β-carboxyl group of each aspartate residue (Simon, 1976, 1987; Simon and Weathers, 1976). Mild hydrolysis of cyanophycin (Joentgen et al., 2001) results in homo- and copolymers of polyaspartate and l-arginine. The copolymer has a broad range of technical and medical applications as an immune system stimulator (Cen et al., 1999; Taheri et al., 2001; de Jonge et al., 2002; Nieves and Langkamp-Henken, 2002; Tapiero et al., 2002), a growth inductor (Roth et al., 1995; Lenis et al., 1999) and a tumour cell growth inhibitor (Flynn et al., 2002). Cyanophycin is synthesized via non-ribosomal polypeptide synthesis in many cyanobacteria (Simon, 1987) and other non-photosynthetic bacteria (Krehenbrink et al., 2002; Ziegler et al., 2002). For cyanophycin synthesis, only one enzyme, the cyanophycin synthetase encoded by cphATe, is necessary to catalyse the ATP-dependent elongation of a cyanophycin primer by the consecutive addition of l-aspartic acid and l-arginine (Ziegler et al., 2002). In cyanobacteria, the polymer is variable in length (25–125 kDa), water-insoluble and stored in membrane-less granules (Allen, 1984; Simon, 1987). Neumann et al. (2005) reported the production of cyanophycin in transgenic tobacco and potato plants expressing the cyanophycin synthetase gene cphATe from Thermosynechococcus elongatus BP-1 under the control of the cauliflower mosaic virus (CaMV) 35S promoter. The maximum amount of cyanophycin produced was 1.14% of dry weight (0.37% of the soluble form and 0.77% of the insoluble form) in tobacco leaves. In the progeny of these plants, the mean content of the polymer was further increased, but the proportion of the soluble form decreased. Loosely packed aggregates of cyanophycin were detected in the cytosol of the leaves. The cyanophycin-producing plants exhibited changes in leaf morphology, such as thickened cell walls, modified chloroplast morphology and slow growth. In the present study, it is shown that transgenic tobacco plants, which constitutively express a nuclear-encoded cyanophycin synthetase directed to the plastids, accumulate higher levels of cyanophycin in the plastids, with minor or no deleterious effects on growth and morphology.

Results

Generation of transgenic N. tabacum plants with plastidic production of cyanophycin

In order to induce plastidic production of cyanophycin, the sequence coding for the cyanophycin synthetase of T. elongatus BP-1 was fused to translocation pathway signal sequences for import into chloroplasts. As the targeting efficiency of transit peptides depends on the proximate coding sequences of the cargo protein (Wong et al., 1992; Jang et al., 2002; Liu et al., 2004), four different translocation pathway signal sequences were utilized. These belong to the following nuclear-encoded proteins: the light-harvesting chlorophyll protein CP24 located in the thylakoid membrane (CP24) (Cai et al., 1993); the ferredoxin-NADP oxidoreductase which is loosely attached to the stromal side of the thylakoid membrane (FNR) (Clausmeyer et al., 1993); the Rieske iron–sulphur protein, a component of the cytochrome b6f complex, located on the luminal side of the thylakoid membrane (Rieske) (Bartling et al., 1990); and the integral protein of photosystem II (PsbY) (Gau et al., 1998). All of these sequences contain the general plastid import sequence and additional sequences for transport to, into or across the thylakoid membrane (Klösgen, 1997). Thus, the utilized constructs contain the neomycin phosphotransferase type II gene (nptII) (Herrera-Estrella et al., 1983), the CaMV 35S RNA promoter, one of the four used translocation pathway signal sequences and the cyanophycin synthetase coding region (cphATe) from T. elongatus BP-1 (Berg et al., 2000). The constructs pCP24-cphATe, pFNR-cphATe, pRieske-cphATe and pPsbY-cphATe (Figure 1) were used for at least two transformation experiments in N. tabacum SRI (Table 1). The plants obtained were named CP24-cphATe, FNR-cphATe, Rieske-cphATe and PsbY-cphATe.

Figure 1.

Map of the nuclear expression vectors carrying plastid target sequences. The coding region of the cyanophycin synthetase gene (cphATe) was fused to one of four transit peptides under the control of the constitutive p35S promoter. The transit peptides Rieske, FNR and CP24 were integrated via NsiI and ClaI restriction. The transit peptide PsbY was integrated via SmaI and NsiI restriction. TP, transit peptides (Rieske, FNR, CP24 and PsbY); p35S, cauliflower mosaic virus (CaMV) 35S promoter; neo, coding region of neomycin phosphotransferase gene; t35S, CaMV 35S terminator; cphATe, coding region of cyanophycin synthetase gene; LB and RB, left and right borders of Agrobacterium tumefaciens binary vector; E, restriction site for EcoRI; B, restriction site for BclI; cyel1-fw/rv, primer for amplifying a specific 856-bp fragment.

Table 1.  Summary of regeneration rate and characterization of regenerated transgenic plants. Nicotiana tabacum transformed with the different vector constructs for constitutive plastidic expression of cyanophycin synthetase, in comparison with the control vector pLH9000
ConstructShoot/explant (mean rate)Regenerated transgenic plantsT0 plants expressing cphATe mRNAT0 plants producing cyanophycinPhenotypic changes of T0 plants*
  • *

    Changes indicate number of regenerated transgenic plants exhibiting morphological modifications, such as flare sections on leaves and early flower induction.

pRieske-cphATe1115 of 80 of 51 of 5
pFNR-cphATe0.91148 of 97 of 87 of 7
pCP24-cphATe0.42115 of 1711 of 1511 of 11
pPsbY-cphATe0.657651 of 7630 of 360 of 30
pLH90001.08160 of 16

Plants transformed with pCP24-cphATe and, to a lesser extent, pPsbY-cphATe showed decreased regeneration frequencies, whereas the other two constructs displayed similar frequencies when compared with the control vector pLH9000 (Hausmann and Toepfer, 1999) carrying only the nptII gene (Table 1).

The transgenic plants FNR-cphATe and CP24-cphATe showed a significantly decreased growth rate. Most of these plants exhibited a modified phenotype consisting of variegated leaves, which were thicker than the leaves of control plants (Table 1, Figure 2). In contrast, four of the five transgenic lines of Rieske-cphATe and all of the PsbY-cphATe lines tested showed no phenotypic changes or growth retardation (Table 1, Figure 2).

Figure 2.

Phenotype of transgenic tobacco lines transformed with either the FNR-cphATe or PsbY-cphATe construct in comparison with the non-transgenic control plant (wild-type). The FNR-cphATe line showed reduced growth and variegated leaves. Only the progeny of PsbY-cphATe line 51 displayed a thicker stem and leaves with strongly prominent leaf veins and smaller anthers in comparison with the other transformants of PsbY-cphATe.

In comparison with the constitutive cytoplasmic p35S-cphATe plants, which showed several short mRNAs (Figure 3a) (Neumann et al., 2005), the presence of the transgene-specific, full-length mRNA of cphATe was demonstrated by Northern blot analysis or reverse transcriptase-polymerase chain reaction (RT-PCR) in the majority of the regenerated plant lines. For instance, in all FNR-cphATe and CP24-cphATe lines that showed the modified phenotype, full-length mRNA of 2.7 kb was detected (Figure 3b–d). Likewise, this full-length transcript was present in five of the eight Rieske-cphATe lines and 51 of the 76 transgenic PsbY-cphATe lines (Table 1, Figure 3b).

Figure 3.

Northern blot analysis of transgenic tobacco lines transformed with p35S-cphATe, PsbY-cphATe, FNR-cphATe and CP24-cphATe primary transformants. Transcript signals were detected with a cphATe-specific probe, and hybridization detected a 2.7-kb transcript. (a) Transcript signals of two plants of the constitutive cytoplasmic p35S-cphATe construct (described in Neumann et al., 2005). (b) Transcript signals of two transgenic PsbY-cphATe tobacco lines. (c) Transcript signals of two transgenic FNR-cphATe tobacco lines. (d) Transcript signals of two transgenic CP24-cphATe tobacco lines. WT, non-transgenic control plant; C, control vector containing the neomycin phosphotransferase type II gene (nptII). The size of the cphA transcript is indicated between the blots.

Analysis of the cyanophycin content of transgenic tobacco plants

The cyanophycin content of each plant was measured twice. Transgenic tobacco plants expressing PsbY-cphATe primarily produced the water-insoluble form of cyanophycin, and only small or undetectable amounts of the soluble form, whereas, in all lines expressing FNR-cphATe or CP24-cphATe, the water-soluble form was detected and the amount of water-soluble cyanophycin in some lines approached the levels seen in the insoluble fraction. No cyanophycin was detected in any of the five transgenic Rieske-cphATe lines tested (Table 2).

Table 2.  Determination of the cyanophycin content and composition (soluble and insoluble forms) in leaves and the phenotype of various transgenic tobacco lines
PlantTransgenic lineContent of cyanophycin (mg/g dry weight)Phenotype
TotalInsolubleSoluble
  1. n.d., not detected; –, no phenotypic changes; +, phenotypic changes, such as flare sections on leaves, early flower induction and slow growth.

Rieske-cphATe  2n.d.n.d.n.d.+
 31n.d.n.d.n.d.
 34n.d.n.d.n.d.
FNR-cphATe 13 2.54 1.660.88+
 18 8.78 3.325.46+
 1910.64 5.844.8+
CP24-cphATe 2913.22 7.046.18+
 38 7.86 4.733.13+
 74 5.43 3.432.0+
PsbY-cphATe 5117.1317.13n.d.
 5816.6416.50.14
 6717.1317.13n.d.
104 8.74 8.74n.d.
10610.210.2n.d.
126 8.52 8.52n.d.
15817.3617.120.24

The polymer was detected in seven of the eight FNR-cphATe and 11 of the 15 CP24-cphATe tobacco lines harbouring the transgene-specific mRNA (Table 1). In FNR-cphATe lines containing the polymer, the mean content of cyanophycin was 6.76 mg/g dry weight and the maximum amount (line 19) was 10.64 mg/g dry weight (Table 2). In CP24-cphATe lines containing the polymer, the mean content was 3.91 mg/g dry weight, with a maximum amount of 13.22 mg/g dry weight in line 29 (Table 2). The plant lines with the highest cyanophycin production showed a strongly decreased growth rate.

The polymer was also present in 30 of the 36 PsbY-cphATe tobacco lines tested. The mean content of cyanophycin in the lines containing the polymer was 8.06 mg/g dry weight, and the maximum amount was found in line 158 (17.36 mg/g dry weight) (Table 2).

Analysis of the descendants of transgenic tobacco plants

All of the tobacco lines tested were fertile and produced seeds. The T1 generation of two FNR-cphATe, two CP24-cphATe and seven PsbY-cphATe lines was analysed further. Seedlings of FNR-cphATe and CP24-cphATe expressing cyanophycin were identified by the occurrence of variegated leaves and a slightly decreased growth rate, bur PsbY-cphATe descendants were indistinguishable from control seedlings. Twenty descendants of each line were analysed for the presence of cphATe by PCR. Cyanophycin was quantified in at least five descendants of each line (Table 3a).

Table 3.  Cyanophycin contents of the T0 generation (and their respective T1 and T2 descendants) of cyanophycin-producing tobacco plants. (a) Total cyanophycin content [mg/g dry weight (dw)] of selected T0 plants expressing gene cphATe, and total and mean cyanophycin contents of their respective T1 descendants. (b) Comparison of the cyanophycin contents in selected T1 descendants of line 51 (total) and their T2 descendants (total and mean)
(a)
PlantT0 plantsT1 generation
DescendantCyanophycin content (mg/g dw)
LineTotal cyanophycin content (mg/g dw)TotalMean
FNR-cphATe18 8.78 1 4.1 
   2 9.0 
   413.5 
   8 4.7 
   9 5.9 
  25 6.89 
  26 2.54 
  27 3.74 
  28 3.74 
  29 7.6 6.17
1910.64 2 5.33 
   3 3.62 
   4 2.18 
   5 0.59 
   6 1.54 
   7 1.64 
   8 2.15 
   9 3.38 
  10 3.89 
  23 4.74 
  25 0.72 
  27 0.89 
  28 1.75 
  30 1.61 2.32
CP24-cphATe2913.22 2 8.7 
   3 7.38 
   6 6.72 
   8 8.26 
  10 7.34 
  23 6.97 
  24 5.62 
  25 4.11 
  26 3.38 
  27 6.74 6.52
 38 7.86 2 3.2 
   4 2.81 
   6 2.37 
  16 1.63 
  27 0.55 
  28 1.31 
  29 1.45 
  30 3.19 2.06
PsbY-cphATe 5117.13 123.54 
   229.04 
   338.39 
   433.33 
   520.5628.97
 5810.47 1 7.99 
   2 2.76 
   3 5.69 
   4 2.84 
   511.35 
   6 3.11 
   7 7.34 
   8 2.81 
   9 6.8 
  10 1.5 5.22
 6717.13 4 2.72 
   5 5.57 
   7 6.16 
   8 5.87 
   9 3.79 4.82
104 8.74 1 5.34 
   2 7.92 
   311.36 
   4 3.62 
   5 6.39 
   6 3.62 
   7 6.91 
   8 5.84 
   916.42 
  10 8.37 7.58
10610.2 1 3.97 
   2 9.25 
   3 3.38 
   418.42 
   5 1.42 7.23
126 8.52 410.91 
   527.39 
   618.38 
  1210.43 
  1313.52 
  14 8.06 
  15 6.6413.62
15817.36 315.23 
   6 7.5 
   720.39 
   810.47 
  11 3.8 
  12 8.77 
  13 7.69 
  14 7.39 
  20 7.010.04
(b)
T1 generationT2 generation
DescendantCyanophycin content (mg/g dw)
Descendant of PsbY-cphATe line 51Total cyanophycin content (mg/g dw)TotalMean
123.54 116.63 
  220.87 
  315.12 
  623.7219.08
229.04 120.75 
  222.65 
  538.2427.21
338.39 148.9 
  268.42 
  351.58 
  459.72 
  543.44 
  764.756.12
433.33 126.8 
  529.53 
  716.61 
 1113.16 
 12 9.01 
 1326.21 
 1421.94 
 1913.87 
 20 9.0718.46
520.56 111.5 
  418.48 
  515.92 
  629.06 
  717.45 
  9 7.74 
 1124.06 
 12 6.62 
 1410.31 
 1517.64 
 16 9.12 
 17 5.53 
 18 7.96 
 19 8.04 
 2010.1113.3

In descendants of FNR-cphATe lines 18 and 19, the mean content of the polymer (line 18, 6.17 mg/g dry weight; line 19, 2.32 mg/g dry weight) was decreased compared with the content in the mother plants (line 18, 8.78 mg/g dry weight; line 19, 10.64 mg/g dry weight). The same was true for descendants of CP24-cphATe lines 29 (13.22 mg/g dry weight in T0; mean content of 6.52 mg/g dry weight in T1) and 38 (7.86 mg/g dry weight in T0; mean content of 2.06 mg/g dry weight in T1).

The T1 progeny of the seven PsbY-cphATe lines differed in their individual water-insoluble cyanophycin contents, but the content of soluble cyanophycin remained nearly stable (up to 0.5 mg/g dry weight). The changes observed for the amount of insoluble cyanophycin were not dependent on the production in T0 plants. In the descendants of T0 plants 58, 67, 106 and 158, the mean cyanophycin level was reduced compared with the T0 plant, although T1 individuals from lines 106 and 158 showed higher cyanophycin contents than the mother plant. This individual increase in some descendants was also observed in T1 descendants of line 104, but the mean cyanophycin content was equal to the parental line. Only T1 descendants of lines 51 and 126 showed an increase in cyanophycin accumulation in the mean of all plant lines tested (Table 3a).

At least three descendants each of five PsbY-cphATe line 51 T1 plants were analysed further (Table 3b). The mean cyanophycin content of all T2 plants analysed remained stable; nevertheless, the mean content of descendants from specific T1 plants differed from the parent plant, leading to an increase or decrease. The progeny of plant 51-3 showed an increase of nearly 100%, even in the mean, and up to 68.42 mg/g dry weight was found in descendant 51-3-2.

Analysis of transgene copy number and cphA RNA steady state level in descendants of PsbY-cphATe plants

In order to investigate the influence of T-DNA integration and copy number of the PsbY-cphATe gene on cphA expression, genomic DNA from T1, T2 and, for line 51, T3 descendants of the PsbY-cphATe lines 51, 67, 104 and 106 was evaluated by Southern blot hybridization (Figure 4a). DNA was digested with BclI and EcoRI, blotted and hybridized to the cphA probe. For all line 51 descendants (32 plants from T1–T3), a single band was observed for BclI- and EcoRI-digested DNA probed with the cphA sequence; this indicates a single-locus T-DNA insertion event. Multiple bands were observed for lines 67, 104 and 106, indicating two to three copies of cphA. Accordingly, the cphA copy number in the descendants differed slightly.

Figure 4.

Analysis of cphA transgene copy number and expression level in T1 and T2 PsbY-cphATe line 51 descendants. (a) Southern blots showing the number of cphATe gene copies integrated in the genome of transgenic PsbY-cphATe tobacco plants (T2 descendants of lines 51, 67, 104 and 106); 50 µg of DNA digested with BclI or EcoRI was probed with the cphATe coding sequence. M, marker. The sizes of the fragments are indicated between the blots. DNA digested with BclI in lane 1 and with EcoRI in lane 2. (b) Northern blot revealing cphA mRNA from T1 and T2 descendants of PsbY-cphATe line 51; samples of 15 µg of total RNA were loaded. M, marker. The sizes of the fragments are indicated on the left side of the blot. Lane 1, 51-3-1; lane 2, 51-3-2; lane 3, 51-5; lane 4, 51-4-11; lane 5, 51-5-17; lane 6, 51-5-18, lane 7, control vector containing neomycin phosphotransferase type II gene (nptII); lane 8, tobacco wild-type. The cyanophycin content of these plants is given in Table 3b.

In order to analyse whether the two transgenes were inherited as a linked marker, additional germination on kanamycin-containing medium was carried out with 150 descendants of seven line 51 T1 plants; 78%–88% of the T2 descendants were kanamycin resistant; hence, a single insertion locus seems to be likely. All resistant plants carried the cphA gene, as shown by PCR (data not shown). These results indicate that the nptII and cphA genes may be inherited together.

Transcript accumulation of cphA was analysed in different descendants of PsbY-cphATe line 51. With actin transcript levels as a control, plants 51-5, 51-4-11, 51-5-17 and 51-5-18 showed the lowest cphA RNA levels (Figure 4b); the highest levels of cphA expression were observed in plants 51-3-1 and 51-3-2. These data correlate with cyanophycin accumulation. The cyanophycin content of these plants is given in Table 3b.

Electron microscopic and immunocytochemical analyses

In electron micrographs, loosely packed cyanophycin aggregates were visible in the cytoplasm of transgenic FNR-cphATe (Figure 5a) and CP24-cphATe lines. In PsbY-cphATe lines, these aggregates were only detected in chloroplasts (Figure 5b). The aggregates were composed of cyanophycin strands, which differed in thickness and length and were arranged in a pile.

Figure 5.

Electron microscopy of cyanophycin in the leaves and roots of tobacco plants. (a–d) In leaves: (a) FNR-cphATe line 19; (b) PsbY-cphATe line 67; (c) PsbY-cphATe line 51-3; (d) immunocytochemical detection of cyanophycin by the anti-cyanophycin antiserum and gold-coupled anti-rabbit immunoglobulin G (IgG) antibody. (e–f) In roots: (e) detection of cyanophycin in the cytoplasm of CP24-cphATe line 38; (f) detection of cyanophycin in proplastids of transgenic PsbY-cphATe line 51-3. (g) Electron microscopy of a yellow leaf of non-transgenic control plants containing chloroplasts. (h) Electron microscopy of a yellow leaf of PsbY-cphATe line 51-3 showing gerontoplasts containing tubular cyanophycin structures. Bars, 1 µm. C, cytoplasm; Ch, chloroplast; Cy, cyanophycin; G, gerontoplasts; P, proplastids; Sg, starch grain; Ts, tubular structure.

In the descendants of the transgenic PsbY-cphATe line 51-3, which contained the largest amount of cyanophycin, the polymer was mainly located at the chloroplast membrane and spread out into the organelle with increasing plant age (Figure 5c).

The aggregates reacted with an antibody raised against cyanobacterial cyanophycin (Figure 5d). After staining plant leaf tissue for protein with the naphthol blue–black dye, the aggregates were visible by light microscopy. When the roots of various transgenic tobacco lines were analysed, cyanophycin was detected in the cytoplasm of FNR-cphATe line 18 and CP24-cphATe line 38 (Figure 5e), whereas the polymer was only present in the proplastids of PsbY-cphATe line 51-3 (Figure 5f).

In older, yellow leaves of line 51, the highest PsbY-cphATe cyanophycin-producing line, the chloroplasts were apparently converted into gerontoplasts during leaf senescence. Interestingly, in these gerontoplasts of the transgenic plants, the granular structures of cyanophycin dissolved, and fibrillar structures were detected (Figure 5h). With increasing degradation of the gerontoplasts, these structures dissolved even further, but still reacted with the antibody against cyanophycin (data not shown). Gerontoplasts were not observed in non-transgenic control (wild-type) plants without cyanophycin (Figure 5g).

Phenotype of cyanophycin-containing tobacco plants

Most lines derived from plants transformed with pFNR-cphATe or pCP24-cphATe were slow growing compared with non-transgenic control tobacco plants, and showed modified phenotypes with variegated, thicker leaves and early flower induction. In contrast, transgenic PsbY-cphATe lines had a phenotype similar to control plants. All cyanophycin-producing plants were fertile. However, line 51 of PsbY-cphATe produced only a few seeds. The same was true for the T1 and T2 generations. Flowers of T1 and T2 plants exhibited a variegated morphology, with anthers shorter than the stigma (Figure 2). Unlike T0 plants, all line 51 descendants displayed thicker stems and leaves with more prominent leaf veins and decreased growth rates. In the lines with short anthers, the number of seeds could not be increased by hand pollination with line 51 pollen, and the pollination of non-transgenic control plants with line 51 pollen also led to decreased seed production.

Discussion

The aim of this study was to establish a competitive crop-based system to produce biodegradable alternatives to petroleum-derived polymers. To date, the degradable polymers produced in plants have included PHB (Nawrath et al., 1994; Houmiel et al., 1999; Bohmert et al., 2000; Poirier and Gruys, 2002; Saruul et al., 2002; Lössl et al., 2003; Menzel et al., 2003; Wrobel et al., 2004; Petrasovits et al., 2007; Purnell et al., 2007) and cyanophycin (Neumann et al., 2005).

Cyanophycin production in transgenic tobacco and potato plants, utilizing the cyanophycin synthetase gene of the cyanobacterium T. elongatus BP-1 (cphATe), was first reported by Neumann et al. (2005), with polymer accumulation up to 1% of dry weight. However, these plants were not fertile and displayed a modified phenotype. As a result of the plastid targeting of the cyanophycin synthetase, reported in this study, stress symptoms were reduced and the accumulation of cyanophycin was increased to more than 6% of dry weight in specific descendants of transgenic tobacco lines. Targeting enzymes to chloroplasts has become a widely used approach to increase the production of heterologous protein in transgenic plants. Various transit peptide sequences, originating from different plant species, have been used to target enzymes to diverse plastid compartments (Clausmeyer et al., 1993; Henry et al., 1994; Bohmert et al., 2000; Hoppmann et al., 2002; Marques et al., 2003).

In this study, four different translocation signal sequences were tested. The signal sequences were isolated from genes of the following nuclear-encoded proteins: (i) the light-harvesting chlorophyll protein CP24, located in the thylakoid membrane (CP24); (ii) the ferredoxin-NADP oxidoreductase, loosely attached to the stromal side of the thylakoid membrane (FNR); (iii) the Rieske iron–sulphur protein, a component of the cytochrome b6f complex, located on the luminal side of the thylakoid membrane (Rieske); and (iv) the integral protein of photosystem II (PsbY). All of these sequences contain the general plastid import sequence and additional sequences for transport to, into or across the thylakoid membrane (Klösgen, 1997). In contrast with our expectations, plants transformed with the constructs pRieske-cphATe, pCP24-cphATe and pFNR-cphATe either did not produce any cyanophycin (pRieske-cphATe) or exhibited an undesired accumulation of the polymer in the cytosol (pCP24-cphATe and pFNR-cphATe). No cyanophycin was detectable in the chloroplast in these lines. The CP24-cphATe and FNR-cphATe plants with cytosolic cyanophycin production showed a significantly decreased growth rate and a phenotype which deviated from that of non-transgenic control plants, and was comparable with the phenotype of plants with cytosolic expression of the cyanophycin synthetase as previously described by Neumann et al. (2005). Hence, it was concluded that neither the CP24 nor FNR transit peptide could target the cyanophycin synthetase to the plastids, at least in this combination. In contrast with the other constructs and in agreement with our expectations, in PsbY-cphATe plants, cyanophycin was exclusively found in plastids. It is presently unclear why only the signal translocation sequence of PsbY results in a transit peptide which allows the translocation of the cyanophycin synthetase into plastids. It is possible that the inability of CP24 and FNR leaders to target the synthetase into plastids is a result of the 12 amino acids introduced by the cloning strategy between the published transit peptide and the core protein. Nevertheless, plastid leader sequences vary dramatically in sequence and size, and it does not seem likely that 12 additional amino acids would have such a pronounced effect. Moreover, it has been shown that the amino acid sequences of passenger proteins have an influence on integration into plastids (Kavanagh et al., 1988; Creissen et al., 1995; Marques et al., 2003). The use of these transit peptides for direction into plastids may lead to a misdirection, irrespective of whether this is caused by the combination of transit peptide and core protein or by the addition of amino acids. Hence, it may be possible that the cyanophycin synthetase protein also has an influence on the plastid integration process.

Only small amounts of cyanophycin (maximum of 1.32% of dry weight) were found in plants transformed with the pCP24-cphATe and pFNR-cphATe constructs, which resulted in polymer production in the cytoplasm (Table 2, Figure 5a,e). Biosynthesis of the polymer in the cell cytoplasm seems to interact negatively with the metabolism of this compartment, resulting in stress symptoms and, possibly, repression of the regeneration of cells producing larger amounts of cyanophycin in the cytoplasm. In contrast, higher levels of cyanophycin were achieved when CphA carried the thylakoid-directed transit peptide PsbY, resulting in a maximum cyanophycin level of 1.73% of dry weight in primary transformants. The polymer accumulated exclusively in chloroplasts of leaves and in proplastids of roots (Figure 5b,c,f).

Thus, higher polymer accumulation in plastids than in cytoplasm is possible. This may be because plastids are descendants of cyanobacteria, in which cyanophycin is synthesized naturally (Borzi, 1887) as a storage substance without causing deleterious effects. Moreover, the metabolites for cyanophycin synthesis are directly available in the chloroplast, as the key enzymes for l-aspartate and l-arginine synthesis are localized in the chloroplast (Forchhammer and Demarsac, 1994; Sugiyama et al., 2004; Slocum, 2005; Chen et al., 2006).

In contrast with cyanobacteria, in which only an insoluble form of cyanophycin is present, in plants both insoluble and water-soluble forms were detected. The ratio of these two forms differed in the plants producing cyanophycin in the cytoplasm or plastids. In transgenic lines with cyanophycin synthesis in the cytoplasm, the mean ratio of the insoluble to water-soluble form was 1 : 0.6, whereas, in transgenic PsbY-cphATe lines with production of cyanophycin in the plastids, the soluble form was absent or only very small amounts were detected: the mean ratio of the insoluble to water-soluble form was 1 : 0.003. The molecular basis for the different solubility properties of cyanophycin is not yet known (Ziegler et al., 2002). Nevertheless, the possibility cannot be excluded that there is a correlation between the presence of the soluble protein and the deleterious effects observed in FNR-cphATe and CP24-cphATe plants.

Constitutive cytosolic production of cyanophycin in p35S-cphATe plants correlates with the formation of several shorter than full-length mRNAs (with sizes between 1000 and 2400 bp), homologous to parts of the cphATe gene (Neumann et al., 2005), and possibly indicating incorrect mRNA processing (Figure 3a). Such an observation has also been made in other publications which have reported the expression of bacterially derived genes in plants (Herrera-Estrella et al., 1983; van Aarssen et al., 1995; Diehn et al., 1998; Haffani et al., 2000). By contrast, the detection of the full-length transcript in FNR-cphATe, CP24-cphATe and PsbY-cphATe plants indicates that the fusion of the cyanophycin synthetase coding region with the transit peptide sequences may enhance mRNA stability. This is in agreement with previously published data showing enhancement of transgene expression by the fusion of various genes with functional chloroplast targeting sequences (Wong et al., 1992; Jang et al., 2002; Liu et al., 2004).

The amount of cyanophycin detected in the individual descendants of transgenic plants varied from line to line, and was either decreased or increased compared with T0 plants. This variability was observed with all three constructs tested. In PsbY-cphATe lines, the greatest increase (up to 6.8% of dry weight) was measured in the T2 generation. There was no tendency for a decrease or increase in the cyanophycin content in the T1 generation compared with the T0 generation. For example, PsbY-cphATe lines 51, 67 and 158 produced the same amount of cyanophycin in the T0 generation (1.7% of dry matter), but, in the T1 progeny, the cyanophycin content increased in line 51 (3.83% of dry matter), whereas, in line 158, the cyanophycin content increased in only one of the nine analysed descendants (2.04% of dry weight) and the cyanophycin content decreased to 0.61% of dry matter in line 67. This phenomenon is consistent with reports in several other publications (Bohmert et al., 2000; Matsumoto et al., 2005; Yang et al., 2005); however, its cause still remains unclear. Many factors can contribute to the variation in transgene expression. These include the integration site, transgene copy number and transgenic locus configuration (Matzke and Matzke, 1998). According to our results, the amount of cyanophycin seems to depend on the number of T-DNA insertion events. The transgenic PsbY-cphATe lines (lines 67, 104 and 106) carrying more than one copy of T-DNA showed lower cyanophycin accumulation than line 51, with only one T-DNA insertion locus. This was consistent with the transcript level and in agreement with studies showing that multiple T-DNA integrations have a negative influence on transcript level (Schubert et al., 2004) and also reduce protein accumulation. In addition, the variability of the polymer content was higher in descendants of plants carrying only one copy of the transgene. The T1 descendants of line 51 showed a higher cyanophycin accumulation, but a high variability in cyanophycin content (standard deviation, 20.31 ± 9.99 mg/g dry weight), in contrast with T1 descendants of line 67 (standard deviation, 4.82 ± 1.49 mg/g dry weight).

In accord with previously published data (Zhang et al., 2007), a correlation was observed between the transcript level and cyanophycin content in descendants of PsbY-cphATe line 51 (Figure 4b): plant 51-3-2, with a higher transcript level of cphA, also showed an increased polymer accumulation of up to 68.42 mg/g dry weight, relative to 51-5-18 (7.96 mg/g dry weight).

The amount of cyanophycin in transgenic plants may also be influenced by cyanophycin degradation. In senescing leaves of transgenic PsbY-cphATe line 51, the structure of cyanophycin changed when chloroplasts transformed to gerontoplasts. The compact granular appearance of cyanophycin aggregates changed to a rather loosely packed structure, which was still detected by the antibody raised against cyanobacterial cyanophycin (data not shown). The complete hydrolysis of proteins to free amino acids depends on the action of endo- and exopeptidases (Brouquisse et al., 2001), but cyanophycin has been shown to be resistant to a range of commercially available proteases (Simon, 1987; Simon and Weathers, 1976). In cyanobacteria, the polymer is degraded by an intracellular exopeptidase called cyanophycinase (Allen, 1984; Simon, 1987), which has not yet been found in plants. This, however, does not rule out the possibility of an unidentified peptidase with broad substrate specificity that functions in plastids, especially gerontoplasts, and may participate in cyanophycin breakdown.

Transgenic PsbY-cphATe line 51 and its progeny produced fewer seeds than non-transgenic control plants. This is probably the result of poor pollen viability, as low seed production also occurred when non-transgenic control plants were fertilized with line 51 pollen. The modified phenotype of short anthers, thicker stems and leaves with strongly prominent leaf veins, observed in all generations and all descendants of self-pollinated line 51, could be caused by somaclonal variations, as the phenotype did not change with increasing cyanophycin content in the descendants, and was only present in this line and not in PsbY-cphATe lines 67 and 158, which showed comparably high cyanophycin levels.

Chloroplast targeting sequences can mediate protein import, not only into chloroplasts, but also into other types of plastid, such as leucoplasts (Boyle et al., 1986; Halpin et al., 1989), etioplasts and amyloplasts (Schindler and Soll, 1986; Strzalka et al., 1987). This suggests that very similar import mechanisms operate in these different types of plastid, and all are able to recognize the same transit peptides. The present work further supports this view, as it was shown that polymer production in PsbY-cphATe lines was detected not only in chloroplasts but also in proplastids of tobacco roots (Figure 5f). In addition, this result strengthens the possibility of increasing cyanophycin production in plant organs such as tubers where no chloroplasts are present.

Our results demonstrate that increased cyanophycin accumulation in transgenic plants is possible, with minor or no stress symptoms, when the cyanophycin synthetase enzyme is directed to the plastids. Moreover, a further increase in polymer accumulation may possibly be achieved by organ-specific expression of the pPsbY-cphATe construct, for example in potato tubers, as the minor stress symptoms observed in line 51 did not increase with increasing cyanophycin content. The utilization of tubers as a production system could lead to cost-effective cyanophycin synthesis in which the polymer is obtained as a by-product of starch isolation.

Eventually, it is hoped that sufficient amounts of cyanophycin may be produced in transgenic plants by further compartmentalization of PsbY-cphATe into amyloplasts of potato tubers and the employment of a strong tuber-specific promoter. During polymer production, the amino acid pool might become limiting. It is likely that this problem could be overcome by influencing specific amino acid synthesis genes.

In recombinant Escherichia coli strains, up to 24% cyanophycin per gram of cellular dry matter can be produced on a technical scale (Frey et al., 2002); however, the need for cost-intensive bioreactors reduces the cost-effectiveness of this production procedure. Therefore, the synthesis of large amounts of cyanophycin in plants represents a possible route for economic production.

Experimental procedures

Growth conditions

Tobacco (N. tabacum) var. Petit Havana SRI was grown on standard Murashige and Skoog (MS) medium (Murashige and Skoog, 1962) with vitamins and 30 g/L sucrose. The plants were cultivated in growth chambers at 24 °C/18 °C with a 16-h/8-h light/dark cycle.

Plasmid construction, plant material and transformation

The transit peptides FNR (Clausmeyer et al., 1993), CP24 (Cai et al., 1993) and Rieske (Bartling et al., 1990) were integrated into the vector p35S cphATe (Neumann et al., 2005) between the p35S promoter and the coding region of cphATe as follows. The transit peptides FNR, CP24 and Rieske were amplified via PCR with sequence-specific primers (Table 4) containing specific restriction sites. These PCR products were digested with NsiI/ClaI and ligated into NsiI/ClaI-digested p35S cphATe. The resulting vectors were named pFNR-cphATe, pCP24-cphATe and pRieske-cphATe.

Table 4.  Sequences of the oligonucleotides used in this study. The restriction sites used are shown in italic type
NameDNA sequence
FNR-fw5′-TACCATGCATATGACCACCGCTGTCACCGCCGCTGTTTC-3′
FNR-rv5′-TACCATCGATTGGGCCCTGATGGGTCCCATTTTCCCA-3′
CP24-fw5′-TACCATGCATATGGCAGCTGCAACATCTGCTACA-3′
CP24-rv5′-TACCATCGATGCAGCTACGACAACCAACTTCCTGC-3′
Rieske-fw5′-TACCATGCATATGATCATCTCCATCTTCAACC-3′
Rieske-rv5′-TACCATCGATATTTGGCATGTCAGCTTCATCCCTCGTA-3′
PsbY-fw5′-TACGCCCGGGATGGCAGCAGCTATGGCAACAGCA-3′
PsbY-rv5′-TACCATGCATGCTGCAGCGAGCTGAGCGATCTGT-TTG-3′
cyel1-fw5′-TTACTGGAGCATTCGGCGTCATA-3′
cyel1-rv5′-TTCCACAATGACGGATTTGGAGA-3′
cphA-fw5′-GTGCCGGCCCATGTGATTGG-3′
cphA-rv5′-AGCCAGGAGCGGCATTGACC-3′
Actin-fw5′-GCAACTGGGATGATATGGAGAA-3′
Actin-rv5′-CTGCCTTTGCAATCCACATCTG-3′

The A. thaliana PsbY transit peptide was amplified with Pfu polymerase using pPsbY1 (Gau et al., 1998). The forward primer PsbY-fw contained a terminal SmaI restriction site and the reverse primer PsbY-rv contained a terminal NsiI site (Table 4). PCR consisted of 28 cycles of 95 °C for 5 min, 60 °C for 30 s and 72 °C for 40 s, with a final extension of 72 °C for 10 min. The PCR product was digested with SmaI/NsiI and integrated into SmaI/NsiI-digested p35S cphATe; the resulting vector was designated pPsbY-cphATe.

The binary vectors were introduced into Agrobacterium tumefaciens LBA4404 (Hoekema et al., 1983). N. tabacum cv. Petit Havana SRI was transformed using Agrobacterium tumefaciens-mediated gene transfer, as described in Wohlleben et al. (1988).

Nucleic acid isolation and analysis

DNA was extracted from 100 mg of plant material using 2V DNA extraction buffer [0.25% sodium dodecylsulphate (SDS), 0.05 m NaOH]. The presence of the transgene in the plant genome was demonstrated using primers (cyel1-fw/rv) (Table 4) which amplify an 856-bp fragment, described by Neumann et al. (2005), located in the coding region of cphATe.

For Southern analysis, genomic DNA was extracted from 3 g of leaf tissue by the cetyltrimethylammonium bromide (CTAB) method (Khanuja et al., 1999). DNA (50 µg) was digested with EcoRI or BclI, separated on 1% agarose gels and transferred to a Hybond N+ membrane (Amersham, Freiburg, Germany). The membranes were prehybridized in SDS-phosphate buffer [7% SDS, 50 mm phosphate buffer, 2% blocking reagent (Roche, Mannheim, Germany), 50% formamide, 5 × saline sodium citrate (SSC), 0.1% laurosylsarcosine] and probed with a cphA PCR fragment of approximately 856 bp produced using the cyel1-fw/rv primers described above. The fragment was labelled using a PCR DIG Probe Synthesis Kit (Roche). Membranes were washed twice at 25 °C with 2 × SSC, 0.1% SDS for 15 min, and then washed twice with 0.1 × SSC, 0.1% SDS at 68 °C for 20 min. Blots were exposed on Kodak Biomax light film (VWR, Darmstadt, Germany).

RNA was extracted from 100 mg of plant leaves as described in Logemann et al. (1987), or using a Qiagen RNeasy Plant Mini-Kit (Hilden, Germany); 15 µg of RNA was separated by 1.5% (w/v) formaldehyde-agarose gel electrophoresis and transferred to Hybond N+ membranes (Amersham) by capillary blotting with 10 × SSC. After UV cross-linking, the membranes were prehybridized, hybridized, washed and detected as described above. Plants were analysed for the expression of the cphA gene in comparison with the expression of the endogenous actin gene from Nicotiana benthamiana (Liu et al., 2005). The actin probe was prepared using the actin forward and reverse primers (Table 4). The fragment was labelled using a PCR DIG Probe Synthesis Kit (Roche).

The RevertAid™ H Minus First Strand cDNA Synthesis Kit (Fermentas, St. Leon-Rot, Germany) was used for cphA gene expression analysis. Each reaction mixture contained 1 µg of DNase-treated RNA, 10 mm dNTP mix, 0.5 µg oligo(dT)-primer, 5 × reaction buffer and 200 U reverse transcriptase, and was incubated at 42 °C for 60 min. The following sequence-specific cphA PCR was performed using the cyel1-fw/rv primers as described above.

Kanamycin germination assay

At least 150 seeds from self-fertilized plants were surface sterilized and placed on Linsmaier and Skoog medium with vitamins, 30 g/L sucrose and 100 mg/L kanamycin. The plates were incubated at 24 °C/18 °C with a 16-h/8-h light/dark cycle. Within 4–5 weeks, green kanamycin-resistant seedlings could be distinguished from white kanamycin-sensitive seedlings.

Analysis of water-soluble and water-insoluble forms of cyanophycin

Both forms of cyanophycin produced in plants were analysed by a modification of the method of Neumann et al. (2005).

Lyophilized plant material (50 mg) was disrupted for 15 min in a bead beater (Retsch MM 301) using three steel beads (2.5 mm in diameter) (Retsch, Haan, Germany) per bin at a frequency of 30 Hz.

The soluble form of cyanophycin was extracted twice with 500 µL of 50 mm Tris/HCl (pH 8.0) by mixing and centrifugation (10 min, 18 000 g), and was purified from the supernatants. The soluble polymer was precipitated with methanol (to 80%).

The pellets were analysed for insoluble cyanophycin as described by Simon (1973) and Ziegler et al. (2002), with some variations. Extraction of the water-insoluble form was performed with HCl in four steps, starting with 400 µL of 0.2 m HCl. The remaining extractions were performed using 0.1 m HCl (300 µL per step).

The cyanophycin content was quantified using the enzymatic procedure described by Neumann et al. (2005).

Microscopy

Electron microscopy, light microscopy and immunocytochemical procedures were performed as described previously (Neumann et al., 2005).

Acknowledgements

We thank Professor R. B. Klösgen (Martin-Luther University of Halle-Wittenberg, Germany) for kindly providing the transit peptides FNR, CP24 and Rieske, and Dr A. Gau (University of Hannover, Germany) for the construct pPsbY1.

We also thank Professor W. Wohlleben (University of Tübingen, Germany) and Dr H. Junghans (NORIKA AG, Gross Lüsewitz, Germany) for many helpful discussions.

This work was supported by the German Federal Ministry of Food, Agriculture and Consumer Protection (BMELV).

Ancillary