Introduction

Polyhydroxyalkanoates (PHA) are polyesters of various hydroxyalkanoates that are synthesized by many gram-positive and gram-negative bacteria. Two important types of PHA are the homopolymer polyhydroxybutyrate (PHB) and the copolymer poly(3-hydroxybutyrate-co-3-hydroxyvalerate) (PHBV). These polymers are accumulated by bacteria up to the level of 90% of dry weight under conditions of nutrient stress and are utilized as a carbon and energy reserve. Commercial production of PHA is hampered due to the high expenses of the fermentation process and polymer substrates (Poirier 2002). Production of PHA on a large agronomic scale in plant systems is seen as the only potential way for large scale production at a lower cost (Houmiel et al. 1999).

Production of PHB homopolymer has been initially reported in transgenic Arabidopsis thaliana (Poirier et al. 1992; Nawrath et al. 1994) and later on was reported in other plants (Slater et al. 1999; Saruul et al. 2002; Lössl et al. 2003), however production of PHBV copolymer has been limited primarily due to the unavailability of the metabolic precursors other than acetyl-CoA (Slater et al. 1999). PHB is synthesized from acetyl-CoA by the sequential action of three enzymes namely, β-ketothiolase (phbA), acetoacetyl-CoA reductase (phbB), and PHB polymerase (phbC) (Fig. 1). Production of PHB in plants is based on the consumption of acetyl-CoA as an initial substrate, therefore oil crops, such as oil palm are potential targets for production of PHB due to their high flux of acetyl-CoA during the oil synthesis stage (Reddy et al. 2003). The oil palm mesocarp is the site for storage oil synthesis, making it an ideal tissue for production of PHB. Introducing the PHB biosynthetic pathway genes into oil palm and mesocarp-specific expression of these genes under the regulation of mesocarp-specific promoters would be desirable and may creates a high potential co-production system to produce PHB with much lower cost. The mesocarp-specific promoter (so-called MSP1; accession no: EU499363) corresponding to the oil palm MT3-A gene that is specifically and abundantly expressed in mesocarp (Siti Nor Akmar et al. 2002; Siti Nor Akmar and Zubaidah 2007) is being used to target the specific expression of PHB genes in mesocarp tissue. Construction of multi-genes vectors carrying PHB and PHBV genes under the regulation of MSP1 has been carried out (Masani and Parveez 2003), and transformation of these vectors into oil palm is still in progress.

Fig. 1
figure 1

Polyhydroxybutyrate (PHB) biosynthetic pathway. PHB is synthesized by the sequential action of β-ketothiolase (phbA), acetoacetyl-CoA reductase (phbB), and PHB polymerase (phbC) in a three-step pathway (Madison and Huisman 1992)

Full size image

Commercial production of PHA in plants requires the development of a well-established transformation technique in order to produce transgenic plants that strongly express the transgene and more importantly in the correct site. Furthermore, production of certain foreign molecules especially pharmaceutical products in plants has recently raised many bio-safety concerns (Mascia and Flavell 2004). Transient expression assays are the rapidly expanding techniques, which have opened up new strategies for transgenic studies without delay and environmental risks associated with the production of stable transgenic plants. In many cases, transient expression systems have been shown to provide ultimately the same information as approaches using transgenic plants (Lu et al. 2007).

In this study, the tissue-specificity and strength of the oil palm mesocarp-specific promoter was initially evaluated via transient assay of GUS reporter gene in different oil palm tissues, including mesocarp, leaf, and root. Subsequently, the expression of PHB biosynthetic pathway genes, driven by the activity of MSP1 was evaluated in transiently transformed mesocarp tissue at both transcriptional and translational levels. This is the first report of using the oil palm mesocarp-specific promoter to regulate expression of foreign genes, and the results showed successful tissue-specific engineering of the entire PHB pathway genes using transiently transformed oil palm tissues as a model system.

Materials and methods

Preparation of target materials for bombardment

Oil palm (Elaeis guineensis, D × P) fruits at 12 weeks after anthesis, young leaves, and roots from 3-month-old geminated seedlings were sterilized in 20% Clorox and rinsed thoroughly with sterile distilled water. Mesocarp part of the fruits and leaves were excised to 1 cm × 1 cm slices. Roots were cut into slices of 1 cm in length. The tissue slices were arranged in the center (approximately 3 cm in diameter) of Petri dishes containing Murashige and Skoog medium (Murashige and Skoog 1962). The tissues were kept at 28°C in a growth chamber 1 day prior to the bombardment.

Plasmid constructs

Plasmid constructs were used as follows: pMS29 plasmid carrying PHB genes (phbA, phbB, and phbC) in fusion to MSP1, constructed by Masani and Parveez (2003), 35SpEGFP and pBI221 plasmids carrying GFP and GUS reporter genes, respectively, in fusion to CaMV35S promoter (Clontech) (Fig. 2). The MSP1pBI221 was constructed by removing the CaMV35S promoter in pBI221 and replacing with MSP1 using HindIII and SmaI restriction enzymes. Plasmid DNA constructs were purified using QIAprep® Miniprep Plasmid Purification Kit (Qiagen).

Fig. 2
figure 2

Schematic diagram of 35SpEGFP (a), pBI221 (b), MSP1pBI221 (c), and pMS29 (d) plasmids. The 35SpEGFP vector was constructed by inserting an 800 bp HindIII-Sma I fragment containing CaMV35S promoter into the MCS of pEGFP-1 (Clontech). The MSP1pBI221 was constructed by replacing the CaMV35S promoter with MSP1. The pMS29 construct was reported by Masani and Parveez (2003); β-ketothiolase: phbA; acetoacetyl-CoA reductase: phbB; PHB polymerase: phbC

Full size image

DNA-microcarrier preparation

Preparation of gold microcarriers was carried out following the instructions of the manufacturer (Bio-Rad Laboratories, Richmond, CA, USA). Sixty milligram of 1-μm diameter gold microcarriers was resuspended in 1 ml of 100% ethanol and vigorously vortexed for 2 min, and then centrifuged at 10,000 rpm for 1 min. The recovered pellet was washed twice with sterile distilled water, and finally the microcarriers were resuspended in 1 ml of sterile distilled water and kept at 4°C.

For each co-bombardment, into an aliquot of 50 μl of microcarriers, 2.5 μl of each DNA constructs (1 μg/μl), 50 μl of CaCl2 (2.5 M) and 20 μl of spermidine (0.1 M) were added one by one followed by continuous vortexing. The mixture was vortexed for an additional 3 min, and then the DNA-microcarriers were recovered by centrifugation at 10,000 rpm for 10 s. The pellet was washed twice with 250 μl of 100% ethanol, and was finally resuspended in 60 μl of 100% ethanol.

Biolistic transformation

Biolistic transformation was carried out by co-bombardment of MSP1pBI221 and 35SpEGFP plasmids into mesocarp, leaf, and root tissues and co-bombardment of pMS29 and 35SpEGFP plasmids into mesocarp and leaf tissues. In a parallel experiment, co-bombardment of pBI221 and 35SpEGFP plasmids was carried out in order to provide a qualitative analysis system for comparison of the CaMV35S and MSP1 promoters. For each co-bombardment, 10 μl of DNA-coated microcarriers was loaded onto the center of macrocarrier and left to be air-dried for almost 2 min. The system was set up based on the optimized parameters for transient transformation of oil palm tissues described by Zubaidah and Siti Nor Akmar (2003). The distances between rupture disk and macrocarrier, macrocarrier and stopping screen, and stopping screen to target were adjusted to 6 mm, 11 mm, and 6 cm, respectively. The bombardment was carried out using rupture disks of 1,550 psi for mesocarp and 1,100 psi for leaf and root tissues while the vacuum was maintained at 27 Hg. The bombarded tissues were kept at 28°C prior to reporter assay.

Histochemical GUS assay in tissues co-bombarded with MSP1pBI221/35SpEGFP gene constructs

The GFP spots were observed in mesocarp, leaf, and root tissue 2 days after co-bombardment, using Nikon SMZ1000 microscope equipped with U.V source and GFP filter (excitation at 360–480 nm and emission at 480–500 nm). Transient expression of GFP reporter gene provided an internal reference for identification of successfully transformed tissues. The GFP-expressing tissue slices were selected for histochemical GUS assay following the procedure described by Jefferson (1987), with some modifications. The transiently transformed mesocarp, leaf, and root tissue slices from each co-bombardment were placed separately in sufficient amount of GUS staining solution (0.1 M NaPO4 buffer, pH 7.0, 10 mM EDTA, pH 7.0, 0.5 mM KFerricyanide, pH 7.0, 0.5 mM KFerrocyanide, pH 7.0, 10 mM X-Gluc, and 0.1% Triton-X) and incubated at 37°C for 12 h. Following GUS staining, the blue spots were observed under the white spectrum using Nikon SMZ1000 microscope.

RT-PCR analysis of PHB genes

GFP analysis was carried out at 12 h and 2 days after co-bombardment of tissues with pMS29 and 35SpEGFP gene constructs. The transiently transformed tissues expressing GFP were used for RT-PCR analysis. The total RNA from non-transformed and transformed mesocarp and leaf tissue slices was extracted according to the protocol described by Gao et al. (2001), and were subjected to RT-PCR analysis. Three sets of primers were designed based on the sequences of the three PHB bacterial genes from the bacterial source of Alcaligenes eutrophus (accession no. J04987) in order to prime the specific amplification of phbA, phbB, and phbC transcripts in RT-PCR reactions (Table 1). In addition gene-specific primers were designed for MT3-A and GFP reporter genes. RT-PCR analysis of MT3-A and GFP reporter genes was also carried out in order to provide internal and positive controls, respectively.

Table 1 List of the primers used in RT-PCR reactions
Full size table

RT-PCR was carried out using the QIAGEN® OneStep RT-PCR Kit following the manufacturer’s instructions. The reactions were set up for each gene separately in a 50-μl reaction mixture, containing 10 μl of 5× QIAGEN OneStep RT-PCR Buffer, 2 μl of 25 mM dNTP, 1.5 μl of each gene-specific forward and reverse primers, 2 μl of QIAGEN OneStep RT-PCR Enzyme Mix, 1 μl of total RNA, and 32 μl of RNase-free water. The RT-PCR cycling was performed as follows: reverse transcription step for 30 min at 50°C, initial PCR activation step for 15 min at 95°C, followed by 30 cycles of denaturation, annealing, and extension steps for 1 min at 94°C, 1 min at 55°C, and 1 min at 72°C, respectively. The final extension step was programmed for 10 min at 72°C. Electrophoresis analysis was carried out on a 1.2% agarose gel at the constant voltage of 80 V for an hour.

TOPO® TA cloning of RT-PCR products and sequence analysis

Cloning of the RT-PCR products was carried out using TOPO TA Cloning Kits (Invitrogen™ life technologies) following the instructions of the manufacturer. Restriction analysis of white colonies (~ten colonies) was carried out using the EcoRI restriction enzyme in order to confirm the presence of the inserts. The recombinant clones were sent for sequencing and the obtained sequences were compared with the sequence of the bacterial PHB genes using the workbench software (http://workbench.sdsc.edu).

SDS-PAGE and western blot analysis of PHB genes

The total protein was extracted from non-transformed and transformed mesocarp and leaf tissue slices according to the protocol described by Teong et al. (2003), and analyzed by separation on a denaturing polyacrylamide gel. Silver staining of the gel was carried using Silver Stain Plus Kit from Bio-Rad. Concentrations of protein samples were determined using Quick Start™ Bradford Protein Assay Kit (Bio-Rad). Blotting of electrophoretically separated protein samples onto Hybond™-C nitrocellulose membrane was carried out using Tris–glycine transfer buffer (25 mM Tris, pH 8.3, 192 mM glycine) at constant current of 400 mA for 2 h using Transphor apparatus (Amersham Biosciences). Blocking of the membrane was carried out for 1 h by soaking the membrane in 1× TBS (20 mM Tris, 500 mM NaCl, pH 7.4) containing 0.1% Tween and 5% dried skimmed milk. Then, the membrane was washed twice for 1 min with fresh changes of washing buffer (0.1% Tween in TBS). Probing of protein blots was carried out using polyclonal primary antibody against phbB and monoclonal primary antibody (Clontech) specific for GFP gene with recommended serial dilutions in TBS. After probing, membrane was washed twice in washing buffer and then the secondary detection was carried out for 1 h by diluting the Goat Anti-Rabbit IgG-AP secondary antibody (Bio-Rad) in TBS. Detection of the signals was carried out using Alkaline Phosphatase Conjugate Substrate Kit (Bio-Rad).

Results

Analysis of tissue-specificity of the MSP1 via histochemical GUS assay

Histochemical GUS assay revealed mesocarp-specific activity of the MSP1 which has directed the expression of GUS reporter gene specifically in mesocarp tissue slices. Panels a1, b1, and c1 in Fig. 3 show the results of GUS assay in transiently transformed GFP-expressing mesocarp, leaf, and root tissue slices (a2, b2, and c2, respectively) 2 days after co-bombardment with MSP1pBI221 and 35SpEGFP gene constructs. As illustrated GUS expression was only detected in mesocarp tissue slices, while no expression level of GUS gene was detected in leaf and root tissue slices. Panels d1, e1, and f1 of the figure show the expression of GUS in GFP-expressing mesocarp, leaf, and root tissue slices (d2, e2, and f2, respectively) co-bombardment with pBI221 and 35SpEGFP gene constructs. As expected, expression of GUS gene was detected in all different tissues, which was driven by the activity of the constitutive CaMV35S promoter. Furthermore no endogenous GUS activity was detected in non-transformed tissues (panels g, h, and i).

Fig. 3
figure 3

Histochemical GUS assay in oil palm mesocarp, leaf, and root tissue slices. Panels a 1 , b 1 , and c 1 : GUS assay in GFP-expressing mesocarp, leaf, and root tissue slices co-bombarded with MSP1pBI221 and 35SpEGFP, respectively. Panels a 2 , b 2 , and c 2 : GFP-expressing mesocarp, leaf, and root tissue slices co-bombarded with MSP1pBI221 and 35SpEGFP, respectively used for GUS assay. Panels d 1 , e 1 , and f 1 : GUS assay in GFP-expressing mesocarp, leaf, and root tissue slices co-bombarded with pBI221 and 35SpEGFP, respectively. Panels d 2 , e 2 , and f 2 : GFP-expressing mesocarp, leaf, and root tissue slices co-bombarded with pBI221 and 35SpEGFP, respectively used for GUS assay. Panels g, h, and i: GUS assay in intact mesocarp, leaf, and root tissue slices, respectively

Full size image

Identification of transiently transformed tissues for analysis of PHB gene expression

The main difficulty of using biolistic-based expression systems is the variability in the number of individual bombardment events between independent experiments. This demands a very large number of assays that need to be performed to obtain statistically meaningful comparisons. However, the use of an internal control, such as GFP reporter gene has been shown to provide a reliable reference for quantitative analysis of transgenes (Schenk et al. 1998). Using GFP reporter gene in co-bombardment with PHB genes provided an estimate of the efficiency of the biolistic delivery and more importantly allowed the selection of transiently transformed tissues which presumably have received both reporter and PHB genes. Furthermore GFP assay being a non-destructive assay allowed subsequent use of the transformed tissues for RT-PCR, SDS-PAGE, and western blot analysis. However, transient expression of GFP gene does not necessarily guaranty the successful expression of PHB genes.

An important consideration was to determine how long after bombardment the gene expression analysis should be carried out in the transiently transformed tissues. This is firstly because that cellular mRNA will be degraded very shortly after translation (Gao et al. 2001) and secondly in certain cases the foreign gene products will be identified as a pathogen or undesirable compounds by the cell defense system which causes the halting of transcription and further exclusion from the metabolic pathways (Lyer et al. 2000). Figure 4 illustrates the transient expression of GFP reporter gene in mesocarp and leaf tissue slices after 12 h and 2 days of co-bombardment with pMS29 and 35SpEGFP gene constructs. Two different time intervals were used for RT-PCR analysis in order to have an indication on the stability and durability of expression of the introduced metabolic pathway.

Fig. 4
figure 4

GFP reporter assay in oil palm mesocarp and leaf tissue slices at different times after co-bombardment with MSP1: PHB and 35S: GFP gene constructs. Panels a and b: GFP expression in mesocarp and leaf tissues, respectively 12 h after co-bombardment. Panels c and d: GFP expression in mesocarp and leaf tissues, respectively 2 days after co-bombardment

Full size image

RT-PCR analysis

Figure 5 shows the result of RT-PCR analysis of PHB genes in mesocarp and leaf tissues, 12 h after co-bombardment with pMS29 and 35SpEGFP gene constructs. RT-PCR analysis showed that the engineering of PHB biosynthetic pathway genes under the regulation of MSP1 in transformed mesocarp tissue has resulted in successful transcription of phbA, phbB, and phbC genes, based on the presence of the expected size bands. The expected sizes of phbA, phbB, and phbC transcripts based on the primers were 261, 173, and 178 bp, respectively. No expression level of PHB genes was detected in both non-transformed and transformed leaf tissues; however RT-PCR analysis of GFP gene (positive control) showed the successful transcription of this gene in transformed leaf tissues, which showed the efficiency and accuracy of both biolistic delivery and RT-PCR techniques.

Fig. 5
figure 5

RT-PCR analysis of PHB genes in non-transformed and transformed mesocarp (a) and leaf tissues (b). Lanes 1 and 2 (RT-PCR of MT3-A in non-transformed and transformed tissues, respectively), lanes 3 and 4 (RT-PCR of GFP in non-transformed and transformed tissues, respectively), lanes 5 and 6 (RT-PCR of phbA in non-transformed and transformed tissues, respectively), lanes 7 and 8 (RT-PCR of phbB in non-transformed and transformed tissues, respectively), and lanes 9 and 10 (RT-PCR of phbC in non-transformed and transformed tissues, respectively)

Full size image

RT-PCR was repeated using total RNA extracted from GFP positive mesocarp and leaf tissues 2 days after bombardment and interestingly produced the same results (data not shown), which indicates that expression of the introduced PHB pathway was stable within 2 days and most likely did not interfere with other metabolic pathways in transiently transformed mesocarp tissues.

Using MT3-A gene as an internal control provided a reliable reference for RT-PCR analysis of PHB genes. It was shown that while RT-PCR was able to detect and amplify the MT3-A mRNA transcript in non-transformed mesocarp tissues, no expression level of PHB genes was detected; indicating that there is no endogenous activity of PHB genes in non-transformed mesocarp tissue and in fact the expression of PHB genes in mesocarp tissue after bombardment is the absolute result of the PHB pathway engineering. All the three PHB transcripts from the RT-PCR step were subjected to sequence analysis in order to confirm their identity. Sequence alignment of the PHB transcripts and the bacterial PHB genes from A. eutrophus showed 100% homologies between sequences.

SDS-PAGE analysis

Comparison of the protein profiles in non-transformed and transformed mesocarp tissues at 2 days after bombardment (Fig. 6) showed the presence of three distinguishable additional bands (nearly 40, 30, and 60 kDa) corresponding most probably to phbA, phbB, and phbC gene products, respectively. The molecular weights of the encoded products of PHB genes are 40, 26, and 60 kDa for phbA, phbB, and phbC, respectively. Furthermore, no distinct changes were observed in the protein profiles from non-transformed and transformed leaf tissues. In order to confirm the above findings western blot analysis was performed using polyclonal antibody specific for phbB gene product.

Fig. 6
figure 6

Analysis of the protein profiles in non-transformed and transformed mesocarp and leaf tissues by denaturating polyacrylamide gene electrophoresis. Panel a: protein profiles in non-transformed (lane 1) and transformed mesocarp tissues (lane 2). Panel b: protein profiles in non-transformed (lane 1) and transformed leaf tissues (lane 2). The additional bands observed with sizes of 40, 30, and 60 kDa, corresponding to phbA, phbB, and phbC protein products

Full size image

Western blot analysis

Due to unavailability of the specific antibody for phbA and phbC gene products, western analysis was only carried out for phbB product and it revealed the successful translation of phbB transcript into protein product in transiently transformed mesocarp tissues 2 days after bombardment (Fig. 7, panel a). Expectedly, due to the mesocarp-specific activity of the MSP1, regulation of phbB gene by this promoter did not result in translation into protein product in transformed leaf tissues (panel c). As illustrated in panels b and d, a band of the expected size of GFP product (55 kDa) was detected in protein blots from transformed mesocarp and leaf tissues challenged with monoclonal antibody specific for GFP product (The GFP protein was not distinguishable in the silver-stained polyacrylamide gel). The reason of including the GFP gene in western analysis was to provide an elite positive control to determine the accuracy and efficiency of blotting and probing procedures.

Fig. 7
figure 7

Western blot analysis of phbB and GFP (positive control) genes in mesocarp and leaf tissues co-bombarded with MSP1: PHB and 35S: GFP. Panel a: western analysis of phbB in blotted protein extracted from transformed and non-transformed mesocarp tissue slices (lanes 1 and 2, respectively). Panel b: western analysis of GFP in blotted protein extracted from transformed and non-transformed mesocarp tissue slices (lanes 1 and 2, respectively). Panel c: western analysis of phbB in blotted protein extracted form transformed and non-transformed leaf tissue slices (lanes 1 and 2, respectively). Panel d: western analysis of GFP in blotted protein extracted from transformed and non-transformed leaf tissue slices (lanes 1 and 2, respectively). The size of phbB and GFP products are 26 and 55 kDa, respectively

Full size image

Discussion

In order to establish a safe strategy for production of transgenic plants, it would be desirable to survey different aspects of the transformation process as well as the expression pattern of the foreign genes prior to stable transformation. This is more crucial when we are dealing with tree crops, such as oil palm where their transgenic production takes several years. The adopted biolistic-based transient assay in this study provided fast and valuable information with respect to the strength and tissue-specificity of the novel oil palm promoter (MSP1), the viability and functionality of the chimeric PHB construct, and the tissue-specific engineering of the PHB biosynthetic pathway genes under the regulation of MSP1.

Constitutive expression of foreign gene products in transgenic plants may be less profitable in comparison to tissue-specific expression, due to the high energy required for such production and problems associated with disposal of the remaining parts (Halfon et al. 1997). Engineering of PHB genes under the regulation of the oil palm MSP1 allows the expression to be channeled specifically in mesocarp tissue.

A prerequisite prior to the use of any promoter in transgenic studies is to characterize the functional behavior of the promoter in the plant that is to express the foreign gene. However, in vivo characterization of plant promoters in perennial crops such as the oil palm through production of transgenic plants is time-consuming. Evaluation of promoter’s activity via transient expression assays provides the possibility to test the functional behavior of the promoter in a short period of time (Schenk et al. 1998).

The present study acknowledges the tissue-specificity of the MSP1, which potentially regulated the expression of PHB genes exclusively in oil palm mesocarp tissue. Mesocarp-specific behavior of MSP1 is the result of the specific interaction between cis-acting promoter elements and mesocarp-specific transcription factors which are produced only in oil palm mesocarp. Apparently such transcription factors are not expressed in oil palm leaves and roots, resulting in failure of detecting GUS reporter gene and PHB transcripts in these tissues.

Several phenomena can effect the expression of the foreign bacterial genes in plant systems, including transcriptional and post-transcriptional gene silencing, DNA methylation of the promoter, codon usage frequency, and base composition of the DNA (Dale and Schantz 2003).

A major concern in using PHB construct for production of transgenic palms was the probability of the occurrence of transcriptional gene silencing due to the repeated use of the same promoter (MSP1) in gene construct. However, the matrix attachment region of tobacco (RB7-MAR) which has a stabilizing effect on transgene expression was incorporated into the PHB gene construct (Masani and Parveez 2003), which might overcome the problem leading to silencing of the transgenes (Allen et al. 1996; Yoshida and Shinmyo 2002). In addition, there was certain degree of uncertainty in transferring PHB genes from bacterial source into a eukaryotic system with different codon usage and the likelihood of slowing down or even halt the transcription due to the codon bias phenomena. The occurrence of codon bias is more prevalent with respect to the genes which have a high expression level, such as genes encoding the enzymes of the metabolic pathways (in our case; PHB genes) (Fisk and Dandekar 1993; Miyasaka 2002).

All these difficulties and uncertainties associated with production of transgenic palms reinforced the need for initial evaluation of the engineering program in a rapid transient biolistic-based assay. The biolistic technique was shown to be a potential approach to test the functional behavior of the MSP1 as well as tissue-specific expression of PHB genes. This study showed the successful transcription of all three PHB genes into mRNA transcripts in transiently transformed oil palm mesocarp tissues. The transcripts remained stable up to 2 days and have successfully been translated into proteins without demonstrating any of the problems associated with expression of multiple foreign genes as discussed earlier. These results are very supportive on the possibility of engineering PHB production in the mesocarp of oil palm. Understanding these data prior to production of transgenic palms brought degrees of certainty on the prospects of long-term transgenic program for PHB-producing transgenic palms.