Cultivated Legume SpeciesView this Special Issue
Research Article | Open Access
Quantitative Trait Loci Analysis of Folate Content in Dry Beans, Phaseolus vulgaris L.
Dry beans (Phaseolus vulgaris L.) contain high levels of folates, yet the level of folate may vary among different genotypes. Folates are essential vitamins and folate deficiencies may lead to a number of health problems. Among the different forms of folates, 5-methyltetrahydrofolate (5MTHF) comprises more than 80% of the total folate in dry beans. The objectives of this paper were to compare selected genotypes of dry beans for the folate content of the dry seeds and to identify quantitative trait loci (QTL) associated with the folate content in a population derived from an inter-gene-pool cross of dry beans. The folate content was examined in three large-seeded (AC Elk, Redhawk, and Taylor) and one medium-seeded (Othello) dry bean genotypes, their six F1 (i.e., one-way diallel crosses), and the F2 of Othello/Redhawk that were evaluated in the field in 2009. Total folate and 5MTHF contents were measured twice with one-hour time interval. The significant variation () in the folate content was observed among the parental genotypes, their F1 progeny, and members of the F2 population, ranging from 147 to 345 μg/100 g. There was a reduction in the 5MTHF and total folate contents in the second compared to the first measurement. Dark red kidney variety Redhawk consistently had the highest and pinto Othello had the lowest total folate and 5MTHF contents in both measurements. A single marker QTL analysis identified three QTL for total folate and 5MTHF contents in the first measurement and one marker for the total folate in the second measurement in the F2. These QTL had significant dominance effects and individually accounted for 7.7% to 10.5% of the total phenotypic variance. The total phenotypic variance explained by the four QTL was 18% for 5MTHF and 19% for total folate in the first measurement, but only 8% for total folate in the second measurement.
Dry beans (Phaseolus vulgaris L.), in addition to being excellent sources of protein and dietary fibre , are also good sources of minerals such as iron, calcium, and zinc, as well as folate . Folate is the general term used to refer to different chemical forms of vitamin B9 . Naturally occurring forms of folate include tetrahydrofolate, 5-methyl tetrahydrofolate (5MTHF), 5-formyltetrahydrofolate, and 10-formyltetrahydrofolate . Among the naturally occurring forms, 5MTHF is the most dominant and readily available form found in plant and animal metabolic cycles . Previous research has indicated that 5MTHF comprises more than 85% of total folate in some major vegetables and close to 100% in some fruits . Hence, in most studies, 5MTHF has been chosen as an indicator for the measurement of total folate present in dry beans. Moreover, 5MTHF is stable in acidic environments, which makes it possible to extract and analyse it with accuracy using High Pressure Liquid Chromatography (HPLC) .
Folates are required in human diets, because humans lack enzymes to synthesize folate de novo . Folate deficiency in humans leads to a number of serious diseases. The occurrence of neural tube defects in infants and different forms of dementia and cardiovascular diseases in adults are due to the deficiency of folate . The current recommended daily dietary allowance of folate is 400 μg for adults and 600 μg for pregnant women and lactating mothers. Folate-rich diets are typically suggested for women planning a pregnancy or who are already pregnant, because of the essential role that folate plays in the production of nucleotides and many other metabolic processes during cell division .
Dry beans generally differ in many traits including seed composition and mineral concentration . Despite high levels of folate in dry beans, limited data may also point to possible differences in the folate content of the genotypes from the two common bean gene pools. This variation may suggest the possibility of improving the folate content of dry beans through plant breeding. The objectives of this paper were, therefore, to (1) compare four varieties of dry beans from the two gene pools, their six F1, and F2 of Othello/Redhawk for the folate content in seeds and (2) identify quantitative trait loci (QTL) associated with the folate content in the F2 of Othello/Redhawk.
2. Materials and Methods
2.1. Plant Materials and Field Trials
Three large-seeded varieties, SVM Taylor Horticulture cranberry bean, henceforth referred to as Taylor, AC Elk light red kidney bean, and Redhawk dark red kidney bean, were selected from the Andean gene pool and the medium-seeded variety, Othello pinto bean, was selected from the Middle America gene-pool. A one-way diallel cross between four parents was made in a growth room at the University of Guelph in the summer of 2009. All six F1 hybrids were grown in the growth room and allowed to produce F2 seeds.
The four dry bean varieties, six F1, and six F2 were grown in the field at the University of Guelph, Elora Research Station, near Elora, ON in the summer of 2009, in a randomized complete block design with three replications. After the harvest and folate measurement, only the F2 of Othello/Redhawk was chosen for further analysis (see below). Each experimental plot of the four varieties and six F1 consisted of a single row, 1.5 meter long with 0.76 meter row spacing. A maximum of 80 seeds of the F2 of Othello/Redhawk were distributed uniformly in six rows, 1.5 meter long with 0.76 meter row spacing. The distance between each plant in the rows was 0.15 meter. At maturity, three plants from each plot of the four varieties, single plants from F1, and every single plant from the F2 of Othello/Redhawk were hand harvested in paper bags and were kept in a dryer at 30°C for 48 hours. Plants were then stored at room temperature until threshed. A single plant thresher was used to thresh one plant at a time. Seeds were collected in separate paper bags. Seeds were cleaned and seed moisture was measured using an Automatic Moisture Meter (Motomco 919E, Paterson, NJ, USA). Dry seeds were stored at −30°C until the folate was extracted.
2.2. Folate Measurement
The folate content was first measured only for four varieties and their six F1. After the data was analyzed, the F2 of Othello/Redhawk was chosen for the genetic analysis, from which every F2 plant was analyzed. The folate content was measured using HPLC with fluorescence detection method at Agriculture and Agri-Food Canada, Guelph Food Research Center (GFRC), Guelph, ON. For each sample, two measurements were taken, with one hour time-interval. Extraction, enzymatic treatment, purification, preparation of standards and HPLC were carried out following the methods explained by Xue et al. .
2.3. Statistical Analyses
Raw data for total folate and 5MTHF in the first and second injections were first compiled and analyzed for four varieties and their six F1. The raw data was subjected to analysis of variance using the PROC GLM procedure of SAS v9.2 , with folate content as the dependent variable and replication and genotype (varieties and F1s) as independent variables. Sums of squares of genotype were partitioned into sums of squares of varieties, crosses, and parents versus crosses. Least square mean values and their standard errors were computed using the LSMEANS statement in PROC GLM. The F2 of Othello/Redhawk was chosen for the genetic study because of highest significant differences between the two varieties (Table 1).
* and ** are significant at P < 0.05 and 0.01, respectively; ns is not significant.|
1In each cross, first parent is the female and second parent is the male.
Statistical analyses of the folate content of the F2 of Othello/Redhawk were conducted using the PROC MIXED procedure in SAS 9.2 using the codes provided by Scott and Milliken  for the Modified Augmented Randomized Complete Block Design, in which varieties and F1 hybrids were the repeated checks and each F2 individual was one experimental unit. To construct the model, two new variables were defined in the data set. The variable C takes the value of a check’s name for checks (varieties and F1s) and the value 0 for experimental units (each F2 single plant). The following model was used: where is the population mean, denote the replication effect, and and denote the entry effect, that is, checks and the F2 individuals. Replications were considered random and checks were considered fixed effects. The nested effect of entry (check) was considered random in the first run of the analysis to generate the Best Linear Unbiased Predictors (BLUP) for each one of the F2 individuals using SOLUTIONS statement in PROC MIXED in SAS  and then considered fixed in a second run of the analysis to generate least square means values for the F2 individuals. The least square means and BLUP values were linearly correlated with a coefficient of correlation greater than 0.99 and, therefore, only least square mean values were used for the QTL analyses.
2.4. DNA Extraction
Leaf tissue samples were taken from young trifoliolate of the four varieties, six F1, and F2 of Othello/Redhawk single plants when the first trifoliolates were fully expanded. Samples were stored on ice in a cooler until transferred to a −80°C freezer. DNA was extracted using a modified FastPrep (Sigma) extraction method following the manufacturer’s protocol. Approximately, one cm2 of frozen leaf, excluding the mid-rib, was placed into a screw-cap tube between a small and a large grinding bead with 600 μL of extraction buffer (200 mM Tris pH 7, 250 mM NaCl, 250 mM EDTA pH 8.0, 0.5% SDS, H2O). The samples were homogenized in a FastPrep (Sigma) grinding machine (Thermo Electron Corporation, Milford, MA, USA) for 20 seconds at 4°C and placed on ice for 5 minutes. The homogenate was pipetted into a new 1.5 mL tube and centrifuged for 5 min at 13200 rpm. An aliquot (400 μL) of the supernatant was transferred to a new 1.5 mL tube and 400 μL of cold isopropanol was added. The tubes were left at room temperature for 5 min and centrifuged for 5 minutes at 13200 rpm. All the supernatants were discarded and the DNA pellets were drained by inversion for 15 minutes and vacuum dried for 10 minutes. Doubled distilled water (500 μL) was added to the samples and was placed at 4°C overnight. The following day, the samples were centrifuged for 1 minute at 13200 rpm. A 480 μL aliquot of the supernatant was collected in a new tube and stored at −20°C until used.
2.5. Molecular Marker Genotyping
The parental lines and the F2 population were genotyped with SNP markers developed by Shi et al. . Genotyping was performed at the Genome Quebec Innovation Center (Montreal, QC) using the Sequenom iPLEX Gold Assay (Sequenom, Cambridge, MA).
2.6. Linkage Mapping
Linkage maps with 67 polymorphic SNP loci were constructed using the software Joinmap 4.0 . Each marker locus was tested for conformity to the expected genotypic ratio of 1 : 2 : 1 in an F2 population using the test in Joinmap 4.0. The markers were first ordered into linkage groups using the “group” command (parameter value LOD > 5). The remaining markers were added to the respective linkage groups based on previous mapping information  using the “assign” command. An alignment comparison of the linkage map was done using the linkage map of McConnell et al.  as a reference map.
2.7. QTL Analysis
A single factor QTL analysis, using one-way ANOVA was used to find the association between polymorphic markers and folate content parameters using PROC GLM in SAS v9.2  with the significant threshold of . The best fit linear model was calculated for each marker and to of the model was computed to estimate the proportion of phenotypic variation explained by each marker. The mathematical model for the single marker QTL analysis was where is the phenotypic value, is the overall mean, is the genotype score of the th marker, and is the residual error. Forward stepwise regression was conducted for significant markers for each trait from single factor QTL analysis to estimate the total phenotypic variance explained by all QTL in the model. The significance of the additive effect at each SNP locus was tested using a single degree of freedom test using ESTIMATE statement in the PROC GLM procedure, in which the phenotypic value of the two homozygous genotypic groups were compared. Similarly, the significance of the dominance effect at each SNP locus was tested with a single degree of freedom contrasts, in which the phenotypic value of the heterozygous genotypic group was compared with the average phenotypic value of the two homozygous genotypic groups. The additive effect at each significant locus was then estimated as half of the difference of the two homozygous genotypic groups and the dominance effect was estimated as the difference of the phenotypic value of the heterozygous group and the mid value of the two homozygous groups.
3.1. Folate Content of Parental Lines and Their F1 Hybrids
Total folate and 5MTHF contents in the solution injected after one hour were significantly different () among genotypes. The four varieties were significantly different () for all four parameters except 5MTHF in the first injected solution (Table 1). Total folate content ranged between 217 and 345 μg/100 g in the first injected solutions and 167 and 321 μg/100 g in the second injected solutions (Table 1). The 5MTHF comprised 70% to 91% of total folate contents in the first injected solution and 72% to 88% of the total folate contents in the second injected solution. Folate contents of Taylor and Othello were significantly lower than AC Elk and Redhawk (Table 1). The reduction of 5MTHF content and total folate content in one-hour time interval between the first and second injections, used as a measure of folate instability, was highly variable for all four parental lines ranging from 5% to 30% for 5MTHF and 7% to 33% for total folate. Othello and Taylor had higher rates of instability of total folate content than Redhawk and AC Elk (Figure 1).
In the analysis of variance, the effect of parent versus crosses was not significant. However, the least square means of the Taylor/Redhawk for all four parameters deviated from mid-parent value towards the high folate content parent (Redhawk) pointing to the involvement of dominance and/or overdominance gene effects in that cross. Similarly, for Taylor/Othello the least square mean values of the F1 for 5MTHF and total folate contents in the first injected solution deviated towards the high folate content parent (Taylor). On the other hand, the 5MTHF and total folate contents for second injected solution of the F1 of Redhawk/AC Elk deviated towards the low folate content parent AC Elk (Table 1).
3.2. Folate Content of the F2 of Othello/Redhawk
The frequency distribution of the folate content for the F2 of Othello/Redhawk was continuous for all four parameters with the presence of transgressive segregants at both ends of the frequency distribution. While the population means for all four parameters were between the values of two varieties, the F1 values were significantly higher than the population means for total folate (both measurements) and were shifted towards the high folate content parent, Redhawk (Figure 2).
The average loss of total folate content in one-hour time intervals from the first injected solution to the second injected solution was 30% for Othello (low folate content parent), but only 7% for Redhawk (high folate content parent). The average loss of folate in one-hour time interval from the first to the second injected solution was 17% for 5MTHF content and 18% for total folate content (Figure 2). Members of the F2 population varied for the proportion of folate lost during the time interval between the first and second injected solutions as evidenced by low, but significant, coefficient of correlation between the folate contents in the first and second injected solutions. The correlations were () for total folate and () for 5MTHF content.
The 5MTHF comprised 70% to 90% of total folate content among the four varieties and F2 individuals. There was highly significant positive correlation between 5MTHF and total folate in the first (; ) and the second (; ) injected solution.
3.3. SNP Analysis and Linkage Map
Among the tested SNP markers, 54% of them were polymorphic between the parents Othello and Redhawk. The total number of SNP markers included in the linkage map was 63, which resulted in a linkage map of 1056.14 cM, in 11 linkage groups (Figure 3), which is equivalent to the haploid chromosome number in the dry bean. The alignment of the linkage map with the map published by McConnell et al.  indicated that with the exception of a mismatch in Pv09 between markers g1286 and g544 all other markers were aligned. However, differences were observed in marker distances.
The chi-square test of conformity of the observed genotypic frequencies for the SNP markers with the expected 1 : 2 : 1 ratio indicated that 15 SNP markers out of a total of 63 markers had significant () segregation distortion. Markers on Pv04 (g755), Pv05 (g1664 and g1883), Pv07 (g1065 and g2357), and Pv10 (g2521_B) were skewed towards Othello and markers on Pv11 (g1438 and g2135) were skewed towards Redhawk. While the segregation distortion of a section of Pv02 (g457 and g680_B) was observed towards Othello, marker g680 was skewed towards Redhawk. Similarly, the segregation distortion of a section of Pv08 (g1084) was observed towards Othello and markers g 2311and g1731 were skewed towards Redhawk.
3.4. Quantitative Trait Loci Analysis
One-way analysis of variance detected a total of four markers significantly () associated with at least one of the folate content measurements (Table 2). These markers are located on Pv02 (g457_B), Pv09 (g1286, g2498), and Pv11 (g2135). Two markers, g1268 and g2498, on Pv09 explained 7.7% and 7.8% of the phenotypic variance, respectively, with a significant () dominance effects for 5MTHF in the first injected solution. These markers (g1268 and g2498) were also significant () for total folate in the first injected solution and explained 7.9% and 7.7% of the phenotypic variance, respectively, with a significant dominance effects. Neither of these markers, however, was significantly associated with measurements taken at the second injected solution. The marker g457_B was the only marker with significant effect on total folate in the second injection, again with significant dominance effect, accounting for 8.1% of the total phenotypic variation. Marker g2135 on Pv11 was significant () for measurements taken at the first injected solution with significant dominance effects, accounting for 9.3% and 10.5% of the variation. The total phenotypic variance explained by the significant markers in a multiple regression model was 18% for 5MTHF and 19% for total folate in the first injected solution, but only 8% for total folate in the second injected solution (Table 2).
|Add: additive effect at each locus estimated as half the difference of the two homozygous genotypic groups. |
Dom: dominance effect estimated as the deviation of heterozygous genotypic group from mid parental genotypes at each locus
: the proportion of phenotypic variance accounted for by each locus.
* and ** are significant at P < 0.05 and 0.01; ns: not significant at P < 0.05.
The results of this study confirmed that despite the high levels of the folate content in dry beans, the genetic variation still exists among different genotypes. Differences of up to 59% for 5MTHF and total folate contents in the first injected solution and up to 92% for 5MTHF content and total folate content in the second injected solution after one hour were observed. Results suggested that the four varieties were not only different for the level of the folate content, but also in its instability and that instability of folate in the extract may have contributed significantly in the variation observed here. Nevertheless, the variation in the folate content and stability in dry beans observed in this study was more pronounced among genotypes from different market classes. The large-seeded Redhawk, AC Elk, and Taylor had a higher folate content than the medium-seeded Othello. Islam et al.  also reported higher concentrations of minerals in the large-seed Andean beans compared with the small- and medium-seeded Middle American beans.
Continuous variation for the folate content, coupled with transgressive segregation, which was observed at both ends of the frequency distribution of folate contents in the F2 of Othello/Redhawk, indicated that quantitative genetic factors are involved in the inheritance of the folate content in dry beans. Blair et al.  reported that the inheritance of iron and zinc accumulation in dry bean seeds was also predominantly quantitative in a recombinant inbred line population of dry beans derived from a cross of DOR364 (low)/G19833 (high). Cichy et al. , however, reported a monogenic inheritance for seed zinc accumulation.
The observation of F2 individuals with folate contents higher than the high folate content parent or lower than the low folate content parent suggests that while the parental lines may share common genes for the folate content, they may have unique loci that interact in the segregating population to determine the folate content in the F2 individuals. Moreover, low but significant correlations between folate content in the first and second measurements indicated that the F2 individuals had a differential response to the one-hour time interval, which may suggest that breeding materials may be different in terms of stability of folate.
Significant marker-QTL associations were identified in this study. Four markers were associated with at least one of the four measurements of the folate content with three markers significantly associated with the folate content in the first injected solution and only one for the solution injected after one hour. The four markers were distributed among three linkage groups. Although most of these QTL did not have large effects individually, together these QTL could control a significant proportion of variation. Furthermore, only dominance effects were significant for these QTL. The small effects of the identified QTL for folate content, the presence of dominance gene effect, and the significant reduction of folate content in the second injected solution as a measure of folate instability, all point to difficulties associated with breeding for a high folate content in dry beans.
The major sources of folate in human diets are legumes, green leafy vegetables, wheat germ, egg yolk, livers, and fortified foods . However, the bioavailability of folate depends on the processing methods and conditions of these sources. A previous study has indicated that large portions of the folate in edible beans are lost during canning processes . A major loss of 5MTHF in navy beans was observed after soaking and cooking processes . Johansson et al.  evaluated the folate content of 10 different precooked vegetarian ready-to-eat meals before and after reheating and reported a significant reduction in the folate content after reheating. In this study, a one-hour time interval between two injections in HPLC was used as a measure of the folate instability in the extract. Among the four varieties, Redhawk and AC Elk had the highest rate of stability in the one-hour time interval. While this may suggest that instability of folate may have contributed significantly in the variation reported here, it may also indicate the need for future research in this area.
Instability of folates in high temperature food processing and therefore folate degradation during cooking and canning processes [10, 18] and not the original level of folate content in dry beans appear to be the major limiting factor in the food produced from dry beans. Results presented here point to the presence of variation in the rate of instability in one hour time interval between the two HPLC injections. Other studies have reported the presence of proteins that bind to folate and protect folate during processing . Protein binding also protects polyglutamyl folates from deglutamylation [21, 22]. Future research on folate content in dry beans should, therefore, be directed towards biochemical mechanisms that may protect the already high levels of folate in the dry beans.
The financial support for this paper was provided by the Ontario Colored Bean Growers’ Association, Agriculture and Agri-Food Canada, University of Guelph, and the Agriculture Adaptation Council of Canada. Technical assistance of Tom Smith, Geoff Worthington, Jan Brazlot, and Chun Shi are gratefully acknowledged.
- M. Duranti, “Grain legume proteins and nutraceutical properties,” Fitoterapia, vol. 77, no. 2, pp. 67–82, 2006.
- M. Hefni, V. Öhrvik, M. Tabekha, and C. Witthöft, “Folate content in foods commonly consumed in Egypt,” Food Chemistry, vol. 121, no. 2, pp. 540–545, 2010.
- J. Arcot and A. Shrestha, “Folate: methods of analysis,” Trends in Food Science and Technology, vol. 16, no. 6-7, pp. 253–266, 2005.
- T. Brody, J. E. Watson, and E. L. R. Stokstad, “Folate pentaglumate and folate hexaglumate mediated one carbon metabolism,” Biochemestry, vol. 21, pp. 276–282, 1982.
- L. Vahteristo, K. Lehikoinen, V. Ollilainen, and P. Varo, “Application of an HPLC assay for the determination of folate derivatives in some vegetables, fruits and berries consumed in Finland,” Food Chemistry, vol. 59, no. 4, pp. 589–597, 1997.
- A. H. Steindal, A. Juzeniene, A. Johnsson, and J. Moan, “Photodegradation of 5-methyltetrahydrofolate: biophysical Aspects,” Photochemistry and Photobiology, vol. 82, no. 6, pp. 1651–1655, 2006.
- S. Roje, “Vitamin B biosynthesis in plants,” Phytochemistry, vol. 68, no. 14, pp. 1904–1921, 2007.
- J. Geisel, “Folic acid and neural tube defects in pregnancy: a review,” Journal of Perinatal Neonatal Nutritions, vol. 17, pp. 268–279, 2003.
- F. M. A. Islam, K. E. Basford, C. Jara, R. J. Redden, and S. Beebe, “Seed compositional and disease resistance differences among gene pools in cultivated common bean,” Genetic Resources and Crop Evolution, vol. 49, no. 3, pp. 285–293, 2002.
- S. Xue, X. Ye, J. Shi et al., “Degradation kinetics of folate (5-methyltetrahydrofolate) in navy beans under various processing conditions,” Food Science and Technology, vol. 44, no. 1, pp. 231–238, 2011.
- SAS Institute, SAS User's Guide, SAS Institute, Cary, NC, USA, 2010.
- R. A. Scott and G. A. Milliken, “A SAS program for analyzing augmented randomized complete-block designs,” Crop Science, vol. 33, no. 4, pp. 865–867, 1993.
- C. Shi, A. Navabi, and K. Yu, “Association mapping of common bacterial blight resistance QTL in Ontario bean breeding populations,” BMC Plant Biology, vol. 11, no. 1, article no. 52, 2011.
- J. W. Van Ooijen, “JoinMap 4, Software for the calculation of genetic linkage maps in experimental populations,” Kyazma B.V., Wageningen, The Netherlands, 2006.
- M. McConnell, S. Mamidi, R. Lee et al., “Syntenic relationships among legumes revealed using a gene-based genetic linkage map of common bean (Phaseolus vulgaris L.),” Theoretical and Applied Genetics, vol. 121, no. 6, pp. 1103–1116, 2010.
- M. W. Blair, C. Astudillo, M. A. Grusak, R. Graham, and S. E. Beebe, “Inheritance of seed iron and zinc concentrations in common bean (Phaseolus vulgaris L.),” Molecular Breeding, vol. 23, no. 2, pp. 197–207, 2009.
- K. A. Cichy, S. Forster, K. F. Grafton, and G. L. Hosfield, “Inheritance of seed zinc accumulation in navy bean,” Crop Science, vol. 45, no. 3, pp. 864–870, 2005.
- J. Dang, J. Arcot, and A. Shrestha, “Folate retention in selected processed legumes,” Food Chemistry, vol. 68, no. 3, pp. 295–298, 2000.
- M. Johansson, C. Furuhagen, W. Frølich, and M. Jägerstad, “Folate content in frozen vegetarian ready meals and folate retention after different reheating methods,” Food Science and Technology, vol. 41, no. 3, pp. 528–536, 2008.
- M. L. Jones and P. F. Nixon, “Tetrahydrofolates are greatly stabilized by binding to bovine milk folate-binding protein,” Journal of Nutrition, vol. 132, no. 9, pp. 2690–2694, 2002.
- G. Orsomando, R. D. De La Garza, B. J. Green et al., “Plant γ-glutamyl hydrolases and folate polyglutamates: characterization, compartmentation, and co-occurrence in vacuoles,” Journal of Biological Chemistry, vol. 280, no. 32, pp. 28877–28884, 2005.
- Y. Wang, Z. Nimec, T. J. Ryan, J. A. Dias, and J. Galivan, “The properties of the secreted γ-glutamyl hydrolases from H35 hepatoma cells,” Biochimica et Biophysica Acta, vol. 1164, no. 3, pp. 227–235, 1993.
Copyright © 2012 S. Khanal et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.