Figure 2. Biosynthesis of gossypol from hemigossypol (Stipanovic et. al. 2005).
Enlarge Button
Figure 2. Biosynthesis of gossypol from hemigossypol (Stipanovic et. al. 2005).

Gossypol is a terpenoid aldehyde compound unique to cotton plants in the Gossypium genus. Cotton plants that contain gossypol are easily recognizable by their black, lysigenous glands on stems and leaves (Figure 1a), and it is also stored in the seed. Of the approximately 50 species in the Gossypium genus, most contain this polyphenolic aldehyde. As seen in figure 2 the compound is derived from hemigossypol through a free radical coupling reaction (Stipanovic et. al. 2005). This coupling reaction produces the two enantiomers (+)-gossypol and (-)-gossypol, depending on the orientation around the binaphthyl bond.

The toxicity of gossypol has made it an area of interest in plant research. When ingested orally, its toxicity can lead to anorexia, severe weight loss, and even death (Eagle 1949). The (-)-gossypol enantiomer will enter the cell and act as an inhibitor of dehydrogenase enzymes involved electron transport (Benz et. al. 1990), which ultimately leads to necrosis of the tissue.

There are also advantageous properties of gossypol to the cotton plant, especially in medicinal uses. (-)-gossypol is a potential aid in cancer research due to its inhibition of cell growth (Band et. al. 1989; Blackstaffe et. all 1997; Shelley et. al. 1999). This is accomplished by reduction of ATP by the uncoupling of oxidative phosphorylation in electron transport (Benz et. al. 1990). HIV research has shown that (-)-gossypol shows anti-HIV-1 activity (Lin et. al. 1993).This enantiomer can be used as an antiamoebic agent (Gonzales-Graza et. al. 1992). There have also been studies in China to suggest its use as a male contraceptive (Matlin et. al. 1985; Lindberg et. al. 1987; Wang et. al. 1987; Yao et. al. 1987).

Research shows that (+)-gossypol is useful to cotton as a natural protection against insects and pathogens. For example, Helicoverpa armigera larvae mature slowly and have a lower survival rate on cotton plants with (+)-gossypol (Yang et. al. 1999). It can also be consumed by nonruminant animals without expressing any signs of toxicity (Stipanovic et. al. 2005).

Cotton is well known as the leading fiber crop in the world (Lusas and Jividen 1987), but it's potential as a food crop has often been overlooked (Alford et. al. 1996). It has actually been classified as the second best potential source for plant proteins and the fifth best oil-producing plant (Texier 1993). Gossypol's toxicity restrains the potential of using cotton as a food crop, which is why emphasis has been placed on breeding for cotton with reduced amounts of gossypol in its seed (Vroh et. al. 1999). With low gossypol in the seed it can be safely used to produce cotton oil.

Materials and Methods

An overview of the process of creating a microarray, from the plant to the microarray slide, is shown in the following PDF: Mic.pdf. RNA is first extracted from the plant tissue, then it is amplified to aRNA and labeled. The labeled aRNA is then hybridized to a microarray slide specific for the cotton genes and the results are analyzed.

Microarray Slides

The microarray slides were designed by Udall et al. in the Wendel Lab at Iowa State University (Udall et al. 2007). Three Gossypium species were used to construct two comprehensive assemblies of cotton ESTs

Glandless and Glanded Cotton Plants

Our first task to accomplish was to grow a population of glanded and glandless cotton plants (Table 1). For the glandless population we used the gl2gl3 species. Twenty seeds were germinated and of those seeds twelve grew to maturation. Two of the glandless plants appeared to be chlorophyll mutants and after further observation expressed glands. As a result they were not used for further experimentation. For the glanded population we used the Siokra L23, TM-1, and AD2 species. Ten Siokra L23 seeds were germinated, six of which grew to maturation; ten TM-1 seeds were germinated, nine of which grew to maturation; and three AD2 seeds were germinated, two of which grew to maturation.

Cotton Accessions
Sample # Description Date Germinated Rep #
MS 47 (A-T) gl2gl3 (Glandless) 5 April 2007 1
MS 48 (A-J) Siokra L23 (Glanded) 5 April 2007 1
MS 49 (A-J) TM-1 (Glanded) 5 April 2007 1
MS 50 (A-C) AD2 (Glanded) 5 April 2007 1
Table 1: Cotton plants that were grown for this experiment.

RNA Extraction

After approximately three months of letting the cotton plants grow we collected samples from each plant. Young leaf tissue was collected, wrapped in tin foil, and stored at -80 °C. An example of tissue samples that were taken is shown in Figure 1 above. It took us a while to work out the protocol for the RNA extractions, but we finally found success in the Hot Borate RNA Extraction Protocol from the Wendel lab in Iowa State University. Results of the RNA extractions are shown in Table 2 with each sample being in a 25 μl solution of RNase-Free water. The blue highlighted samples in the table were used for RNA amplification to aRNA.

RNA Amplification to aRNA

After extracting RNA from each sample it was diluted down to concentrations of approximately 1,000 ng/μl in 10μl (see Table 3). The samples were then grouped into pools for amplification. There were a total of six groups: three glandless (gl2gl3), and three glanded (TM-1). The highlighted samples in Table 3 show the samples that were pooled together (i.e. MS 51, MS 52, and MS 53 were grouped together in one pool). We wanted a concentration of approximately 1,000 ng/μl in each of the grouped samples with an equal amount of RNA from each individual sample. The concentrations of each newly grouped pool are shown in Table 4.

After grouping the RNA samples into their respective pools they were amplified. The Epicentre Biotechnoligies TargetAmp 1-Round aRNA Amplification Kit 103 (Original Protocol here) was used for the amplification. Results of the aRNA amplification are shown in Table 5 with each sample being in a 50 μl solution of RNase-Free Water.

aRNA Labeling

The next step in the procedure was labeling the aRNA.

Slide Hybridization

Microarray Analysis


The glandless gl2gl3 population and the glanded TM-1 population were chosen for comparative analysis through microarrays. The TM-1 population was chosen because it showed the closest phenotypic resemblance to the gl2gl3 population (i.e. they were about the same height and their leaves looked similar in size and shape). This was done in hopes to narrow down the options of the expression of genes involved in the gossypol synthesis pathway.

After several attempts at various protocols we finally found one protocol that gave us good results for the RNA extractions. As seen in table 2 we extracted anywhere from 1.8 µg to 4.3 µg of RNA. The blue highlighted are the gl2gl3 and TM-1 plants chosen in this experiment to be used for microarrays. Sample MS 47 D and MS 47 L were not chosen because they appeared to be chlorophyll mutants and also showed glands on the leaves and stems when they were supposed to be glandless. The samples were then diluted down to about 1000 ng/µl to be used as templates for amplification. The results of the dilutions are shown in table 3. The remaining stock solution of RNA was stored at -80 °C to be used for further experimentation.

Table 2 
Extracted RNA Concentrations (25 µl)
Blue samples were diluted for microarrays
Sample # Accession Description ng/µl
MS 47 A gl2gl3 Glandless 3205.8
MS 47 B gl2gl3 Glandless 2926.6
MS 47 C gl2gl3 Glandless 3484.8
MS 47 D gl2gl3 Glanded 2623.7
MS 47 E gl2gl3 Glandless 2777.7
MS 47 F gl2gl3 Glandless 2440.3
MS 47 G gl2gl3 Glandless 2204.7
MS 47 H gl2gl3 Glandless 2301.8
MS 47 I gl2gl3 Glandless 2526.0
MS 47 J gl2gl3 Glandless 2110.9
MS 47 K gl2gl3 Glandless 2080.3
MS 47 L gl2gl3 Glanded 1916.4
MS 48 A Siokra L23 Glanded 4299.2
MS 48 B Siokra L23 Glanded 3581.7
MS 48 C Siokra L23 Glanded 3804.6
MS 48 D Siokra L23 Glanded 2815.6
MS 48 E Siokra L23 Glanded 3045.6
MS 48 F Siokra L23 Glanded 2888.3
MS 49 A TM-1 Glanded 3297.1
MS 49 B TM-1 Glanded 2952.0
MS 49 C TM-1 Glanded 3214.7
MS 49 D TM-1 Glanded 1170.7
MS 49 E TM-1 Glanded 3407.3
MS 49 F TM-1 Glanded 3690.2
MS 49 G TM-1 Glanded 2713.1
MS 49 H TM-1 Glanded 3126.5
MS 49 I TM-1 Glanded 2271.4
MS 50 A AD2 P-57 Glanded 1850.4
MS 50 B AD2 P-57 Glanded 2261.3
Table 3 
Diluted RNA Concentrations (10 µl) 

Each color was grouped for microarrays
Sample # Accession Description ng/µl
MS 51 gl2gl3 Glandless 927.9
MS 52 gl2gl3 Glandless 1010.2
MS 53 gl2gl3 Glandless 1013.4
MS 47 D gl2gl3 Glanded 0
MS 54 gl2gl3 Glandless 869.1
MS 55 gl2gl3 Glandless 1010.1
MS 56 gl2gl3 Glandless 930.4
MS 57 gl2gl3 Glandless 945.7
MS 58 gl2gl3 Glandless 870.5
MS 59 gl2gl3 Glandless 971.5
MS 60 gl2gl3 Glandless 1090.4
MS 47 L gl2gl3 Glanded 0
MS 48 A Siokra L23 Glanded 0
MS 48 B Siokra L23 Glanded 0
MS 48 C Siokra L23 Glanded 0
MS 48 D Siokra L23 Glanded 0
MS 48 E Siokra L23 Glanded 0
MS 48 F Siokra L23 Glanded 0
MS 61 TM-1 Glanded 1100.7
MS 62 TM-1 Glanded 1035.6
MS 63 TM-1 Glanded 1006.2
MS 64 TM-1 Glanded 974.8
MS 65 TM-1 Glanded 926.7
MS 66 TM-1 Glanded 970.0
MS 67 TM-1 Glanded 955.6
MS 68 TM-1 Glanded 973.2
MS 69 TM-1 Glanded 912.8
MS 50 A AD2 P-57 Glanded 0
MS 50 B AD2 P-57 Glanded 0

The diluted samples were pooled together in groups of three (and one case of a group of four in the gl2gl3 population) as seen by the colored groups in table 3. In each of the six total pools we wanted an equal amount of the individual RNA samples and our target concentration was 1000 ng/µl. By pooling them together in this way we expected to see an equal amount of gene expression from each plant. As seen from the results in table 4 the samples were all close to the target concentration.

From each of the newly formed pools we used 500 ng of RNA to be amplified. The results of the amplification are shown in table 5. Although we expected better concentrations from the RNA amplification, we still had enough aRNA to continue with the microarrays. Based on the spectrophotometry analysis of the aRNA the samples were clean and showed no sign of contamination. The amplified RNA concentrations were possibly lower than expected due to incubation inconsistencies (i.e. the heating block used for incubation dropped below the desired incubation temperature).

Three microarray slides were prepared for the analysis. On each slide we hybridized on pooled sample of gl2gl3 and one pooled sample of TM-1. The glandless sample was labeled with one dye and the glanded sample was labeled with the other dye. By doing so we should be able to see the expression of genes involved in the gossypol pathway when there is a distinct expression of the glanded dye and not the glandless dye. Cy3 and Cy5 were used to label the samples. On our first slide MS 70 (glandless) was labeled with Cy3 and MS 73 (glanded) was labeled with Cy5. We also labeled MS 72 (glandless) with Cy3 and MS 75 (glanded) with Cy5 for the third microarray slide. For the second microarray slide we flipped the label on the samples. In other words MS 71 (glandless) was labeled with Cy5 and MS 74 (glanded) was labeled with Cy3. We flipped the dyes to check that the labeling and hybridization was done correctly. In the resulting microarray analysis slide 1 and 3 would show glanded gene expression as red "spots" and slide 2 would show glanded gene expression as green "spots."

The six pooled samples were labeled with their respective dyes as was just previously described. To check the quality of labeling a frequency of incorporation (FOI) was calculated. Table 6 shows each labeled sample with its FOI. The optimal FOI would be at least 20, but we are able to work with anything over an FOI of 10. We are assuming that there was a lower FOI because the labeling kit we used was close to its expiration. As seen from the results of the microarrays the label and its FOI was sufficient for analysis.

Table 4 
Combined RNA Concentration
Each sample is a combination of the diluted samples. 
The target concentration was 1000 ng/µl
Sample # Accession Samples Combined ng/µl
MS 70 Combined gl2gl3 (Glandless) MS 51, 52, 53 1043.5
MS 71 Combined gl2gl3 (Glandless) MS 54, 55, 56 1016.7
MS 72 Combined gl2gl3 (Glandless) MS 57, 58, 59, 60 1046.3
MS 73 Combined gl2gl3 (Glanded) MS 61, 62, 63 1079.6
MS 74 Combined gl2gl3 (Glanded) MS 64, 65, 66 1066.8
MS 75 Combined gl2gl3 (Glanded) MS 67, 68, 69 1009.9
Table 5 
Amplified aRNA Concentration (50 µl)

Each combined sample was amplified to make aRNA.
Sample # Accession Samples Combined ng/µl
MS 70 aRNA gl2gl3 (Glandless) MS 51, 52, 53 71.5
MS 71 aRNA gl2gl3 (Glandless) MS 54, 55, 56 101.8
MS 72 aRNA gl2gl3 (Glandless) MS 57, 58, 59, 60 93.4
MS 73 aRNA gl2gl3 (Glanded) MS 61, 62, 63 89.1
MS 74 aRNA gl2gl3 (Glanded) MS 64, 65, 66 51.9
MS 75 aRNA gl2gl3 (Glanded) MS 67, 68, 69 54.8
Table 6 
aRNA Frequency of Incorporation (FOI)
Each aRNA sample was labeled with Cy5 or Cy3 for hybridization preparation.
Sample # Accession Label Type FOI
MS 70 Cy3 gl2gl3 (Glandless) Cy3 13.83
MS 71 Cy5 gl2gl3 (Glandless) Cy5 12.85
MS 72 Cy3 gl2gl3 (Glandless) Cy3 17.38
MS 73 Cy5 gl2gl3 (Glanded) Cy5 15.10
MS 74 Cy3 gl2gl3 (Glanded) Cy3 19.95
MS 75 Cy5 gl2gl3 (Glanded) Cy5 10.48

After the samples were labeled we pooled one glandless sample with one glanded sample (see table 7). After their preparation they were hybridized to the microarray slides. The prepared slides were then scanned. Results of each microarray slide are shown in the links in table 7. A microarray hybridization record was also made to keep track of the microarrays that have been done.

Table 7 
Microarray Results
Slide # Samples Hybridized Image of Microarray Computer Analysis of Microarray
Slide 1 MS 70/MS 73 MS 70/MS 73 File:Analysis Slide 1.pdf
Slide 2 MS 71/MS 74 MS 71/MS 74 File:Analysis Slide 2.pdf
Slide 3 MS 72/MS 75 MS 72/MS 75 File:Analysis Slide 3.pdf

After the slides were scanned they were prepared for computer analysis. Each slide was aligned with a grid for the computer to analyze the expression of each gene individually. The computer analysis is provided in the links from table 7.



  • Alford BB, Liepa GU, Vanbeber AD (1996) Cottonseed protein: what does the future hold? Plant Food Hum Nutr 49:1-11
  • Band V, Hoffer AP, Bands H, Rhinehardt AE, Knapp RC, Matlin SA, Anderson DJ (1989) Antiproliferative effect of gossypol and its optical isomers on human reproductive cancer cell lines. Gynecol Oncol 32:273-277
  • Benz CC, Keniry MA, Ford JM, Townsend AJ, Cox FW, Palayoor S, Matlin SA, Hait WN, Kowan KH (1990) Biochemical correlates of the antitumor and antimitochondrial properties of gossypol enantiomers. Mol Pharmacol 37:840-847
  • Blackstaffe L, Shelley MD, Fish RG (1997) Cytotoxicity of gossypol enantiomers and its quinone metabolite gossypolone in melanoma cell lines. Melanoma Res 7:364-372
  • Eagle E (1949) Chronic toxicity of gossypol. Science 190:361
  • Gonzales-Garza MT, Matlin SA, Mata-Cardenas BD, Said-Fernandez S (1992) Further studies on the in vitro activity of gossypol as antiamebic agent. Arch Med Res 23:69-70
  • Lin TS, Schinazi RF, Zhu J, Birks E, Carbone R, Si Y, Wu K, Huang L, Prusoff WH (1993) Anti-HIV-1 activity and cellular pharmacology of various analogs of gossypol. Biochem Pharmacol 46:251-255
  • Lindberg MC, Naqvi RH, Matlin SA, Zhou RH, Bialy G, Blye RP (1987) Comparative anti-fertility effects of gossypol enantiomers in male hamsters. Int J Androl 10:619-623
  • Lusas EW, Jividen EM (1987) Glandless cottonseed: a review of the first 25 years of processing and utilization research. J Am Oil Chem Soc 64:839-854
  • Matlin SA, Zhou R, Bialy G, Blye RP, Naqvi RH, Lindberg MC (1985) (-)-Gossypol: an active male antifertility agent. 31:141-149
  • Shelley MD, Hartley L, Fish RG, Groundwater P, Morgan JJ, Mort D, Mason M, Evans A (1999) Stereospecific cytotoxic effects of gossypol enantiomers and gossypolone in tumor cell lines. Cancer Lett 135:171-180
  • Stipanovic RD, Puckhaber LS, Bell AA, Percival AE, Jacobs J (2005) Occurrence of (+)- and (-)-gossypol in wild species of cotton and in Gossypium hirsutum var. marie-galante (Watt) Hutchison. J Agric Food Chem 53:6266-6271
  • Texier PH (1993) Le coton, cinquieme producteur mondial d'huile alimentaire. Cotton Dev 8:2-3
  • Vroh BI, Maquet A, Baudoin JP, du Jardin P, Jacquemin JM, Mergeai G (1999) Breeding for "low-gossypol seed and high-gossypol plants" in upland cotton. Analysis of tri-species hybrids and backcross progenies using AFLP's and mapped RFLP's. Theor Appl Genet 99:1233-1244
  • Wang NG, Zhou LF, Guan MZ, Lei HP, (1987) Effects of (-)- and (+)-gossypol on fertility in male rats. J Ethnopharmacol 20:21-24
  • Yang WH, Ma LH, Zhu HQ, Xiang SK, (1999) Effects of different gossypol enantiomers on growth and development of cotton bollworm (Helicoverpa armigera) andFusarium wilt. Acta Gossypii Sinica 11:31-34
  • Yao K, Gu Q, Lei H, (1987) Effect of (+/-)-, (+)-, and (-)-gossypol on the lactate dehydrogenase-X activity of rat testis. J Ethnopharmacol 20:25-29