Identification of factors for improved ethylene production via the ethylene forming enzyme in chemostat cultures of Saccharomyces cerevisiae
© Johansson et al.; licensee BioMed Central Ltd. 2013
Received: 23 July 2013
Accepted: 30 September 2013
Published: 1 October 2013
Biotechnological production of the traditional petrochemical ethylene is presently being explored using yeasts as well as bacteria. In this study we quantify the specific ethylene production levels at different conditions in continuous (chemostat) cultivation of Saccharomyces cerevisae expressing the ethylene forming enzyme (EFE) from Pseudomonas syringae.
Our study shows that oxygen availability is an important factor for the ethylene formation. Maintaining a high percentage dissolved oxygen in the cultivation was found to be necessary to achieve maximal ethylene productivity. Even at oxygen levels high enough to sustain respiratory metabolism the ethylene formation was restricted. Oxygen was also important for sustaining a high respiratory rate and to re-oxidize the surplus of NADH that accompanies ethylene formation. By employing three different nitrogen sources we further found that the nitrogen source available can both improve and impair the ethylene productivity. Contrary to findings in batch cultures, using glutamate did not give a significant increase in specific ethylene production levels compared to the reference condition with ammonia, whereas a combination of glutamate and arginine resulted in a strongly diminished specific ethylene production. Furthermore, from cultivations at different dilution rates the ethylene formation was found to be coupled to growth rate.
To optimize the ethylene productivity in S. cerevisiae expressing a bacterial ethylene forming enzyme, controlling the oxygen availability and growth rate as well as employing an ideal nitrogen source is of importance. The effects of these factors as studied here provide a basis for an optimized process for ethylene production in S. cerevisiae.
KeywordsYeast Saccharomyces cerevisiae Ethylene Chemostat Productivity
Despite a general consensus that the dependency on finite resources such as oil and natural gases should be reduced, the consumption rate of these continues to increase. The main usage of these resources is as fuels. However, a wide range of other chemicals are also produced from petroleum, among which the most abundant are hydrocarbon monomers such as ethylene (ethene, C2H4). Due to an increased usage in foremost Asia, the demand on ethylene has increased with an average of 2.5% per year in the last decades and in 2012 the global demand for ethylene was roughly 124 million tons per year[2–4]. The reason for this demand is the versatility of ethylene as a building block for a wide range of compounds, primarily its polymer product polyethylene which forms the basis for many plastic materials.
Biological production of ethylene could be an alternative to the traditional petroleum based chemical method. Several types of microorganisms, including both bacteria and fungi, have been reported to naturally produce ethylene. Two different pathways for ethylene production have been identified within these organisms. One group produces ethylene via the KMBA pathway in which methionine is converted into ethylene in a two-step process, whereas the other group employs a single enzyme, the ethylene forming enzyme (EFE), to convert 2-oxoglutarate directly into ethylene[6, 7]. Measuring the ethylene productivity of 757 bacterial strains, Pseudomonas syringae pv. Phaseolicola PK2 was found to be the most productive species. This strain was classified as belonging to the 2-oxoglutarate dependent group.
The major advantage of utilizing microorganisms for the production of chemicals is the possibility to exploit biomass, such as forest and agricultural residues, as raw material in the fermentative process. Biomass is a plentiful and renewable material and microorganisms can, via metabolic engineering, be altered to acquire the ability to transform the carbons (mainly sugars) in the biomass into desired products as reviewed by e.g. Nielsen et al. and Zhang et al.. On this basis, several biological systems for the conversion of sugars into ethylene have been examined. Heterologous ethylene production via mainly the P. syringae EFE has been reported for a variety of fungi[10, 11] and bacteria[12–14]. A further alternative for biological ethylene production that has been investigated is the conversion of sunlight and CO2 into ethylene via cyanobacteria[15, 16].
Some initial cultivation studies of these ethylene producers have been performed, primarily in shake flask and batch setup in bioreactors. Several of these studies have identified factors which could be restricting the ethylene reaction and hence negatively affect the ethylene yields. The most commonly mentioned factors are; oxygen availability[12, 17], respiration rate and substrate provision[10, 12, 16]. Since optimal production levels are required to make biological ethylene production a feasible alternative to the traditional petroleum based production method, each of these limitations must be studied to understand the extent to which they affect ethylene yield and thereby identify approaches for improving the ethylene production ability.
Using EFE expressed in the yeast S. cerevisiae, this study addresses the most commonly indicated limiting factors of biological ethylene production. A strain of the yeast S. cerevisiae engineered to express the P. syringae EFE was employed. S. cerevisiae was chosen due to the ease with which genetic manipulation can be performed in it, but also due to its relative sturdiness and the fact that it is already used as a biological factory in industrial settings. Cultivations were performed in a chemostat mode to facilitate the comparison of different conditions. Oxygen level, respiration rate and substrate provision were all addressed in different experiments. Furthermore, strategies to increase respiration rate were evaluated for their effect on ethylene production levels.
Results and discussion
Theoretical modeling of this reaction together with the central carbon metabolism has proposed routes for improved ethylene production. In this study we take into consideration a number of the factors suggested to affect the ethylene productivity. Each of the proposed substrates of EFE – 2-oxoglutarate, arginine and oxygen - have been studied with respect to their effects on ethylene formation using well defined chemostat conditions. We further investigated the effect of dilution rate and dissolved oxygen levels as well as the respiration rate.
The effect of nitrogen source on ethylene production
The effect of nitrogen source on the ethylene production
Biomass [g L-1]
Productivity [μg LCulture-1h-1]
Specific productivity[μg gDW-1h-1]
Yield [μg gGlucose-1]
(NH4)2SO4 (7.5 g/L)
5.72 ± 0.95
178 ± 25
30.4 ± 2.8
164 ± 21
Glutamate (7.5 g/L)
7.48 ± 0.14
242 ± 2
32.3 ± 0.3
233 ± 0.8
Glutamate + Arginine (3.5 g/L each)
7.34 ± 0.03
101 ± 1
13.8 ± 0.2
96.8 ± 1.0
The two reactions have been reported to occur at an average ratio of 2:1, using purified enzyme. It is possible that by supplying extra arginine we could be pushing the balance between the two reactions towards the non-ethylene forming reaction (reaction 3), in which arginine is a substrate instead of a co-factor. Hence, less ethylene would be formed per glucose which would in turn explain the decreased ethylene formation found in this study when combining arginine and glutamate as nitrogen source.
Effect of oxygen availability on ethylene production
Fukuda et al. identified oxygen as an essential factor for the ethylene formation in vitro using purified EFE. Several later studies have also proposed oxygen provision as a limiting factor for ethylene formation via the EFE in vivo[10, 12, 18]. To elucidate the effect of oxygen availability on the ethylene formation via the EFE we also investigated the effect of dissolved oxygen concentration on the ethylene productivity.
It was found that an alteration in dissolved oxygen levels was followed by a relatively fast response (within the interval of sampling, which was one hour for this experiment), in which an increase in oxygen level was followed by an elevation of specific ethylene production, and a decrease in oxygen level was followed by a reduction in ethylene production.
Previous studies of continuous cultures of S. cerevsiae have indicated that even very low percentages of oxygen can sustain full respiration at a dilution rate of 0.1 h-1[25, 26]. Accordingly, the biomass formation was found to be constant over the full range of dissolved oxygen percentages applied. Furthermore, no significant production of ethanol or glycerol could be detected. Therefore, the effect on ethylene formation found in this study can be coupled directly to the oxygen availability for the enzyme rather than a major metabolic shift of the yeast. This indicates that the affinity of EFE for oxygen is low and a good oxygen provision must therefore be ensured for optimal ethylene production. This has implications for how cultivations should be set up, especially when performing studies of ethylene production in yeast strains using shake flasks, where oxygen levels will decrease as cell densities increase. In this case, care should therefore be taken to either ensure proper aeration or to work at low cell densities.
The effect of growth rate and respiration rate on ethylene production
In general, the specific ethylene production rate (μgethylene gDW-1 h-1) increased with increased dilution rate, as can be seen in Figure 2A. Between the dilution rates of 0.25 h-1 and 0.3 h-1 a noticeable jump in specific productivity can be seen. At this point the yeast switches to respiro-fermentative growth. Indicative of this switch is the decrease of biomass formation and the onset of ethanol formation, both of which can be seen for the cultivations (Figure 2B). A further indicator of the shift to respiro-fermentative growth is the overflow of pyruvate from the glycolysis which was also noticed in these cultures. As the dilution rate was increased to 0.30 h-1 there was, moreover, a sharp increase in extracellular acetate concentration, also typical for the shift to partly fermentative growth in S. cerevisiae. This point of respiro-fermentative onset correlates with previous found values for S. cerevisiae[29, 30], indicating that the EFE does not impose any drastic metabolic burden on the cell.
Furthermore, each increase in dilution rate was linked to an increase in specific glucose uptake rate (data not shown). As the glucose uptake rate increased more than the ethylene productivity for each increase in dilution rate, the ethylene yield (μgethylene gglucose-1) declined over the dilution rate span employed (Figure 2A).
The effect of respiration rate on ethylene formation
Specific productivity [μg gDW-1h-1]
Productivity [μg LCulture-1h-1]
30.4 ± 2.8
178 ± 25
+ 7.5 mM Benzoate
50.3 ± 1.3
37.7 ± 0.9
+ 1 mM Azide
To further investigate the correlation between respiration and ethylene formation, we performed cultivations with sodium azide, which effectively blocks the respiration. When azide was added to the culture ethylene formation was immediately prohibited, which is a further indication of the close correlation between respiratory rate and ethylene formation in S. cerevsiae.
The results presented above suggested that increasing the respiratory rate of S. cerevisiae should result in higher ethylene yields, however the increased respiration rate should not come at too high an expense for the cell. One alternative could be to use the TM6* strain of S. cerevisiae which contains a chimeric point mutated version of the Hxt1 and Hxt7 glucose transporters. This strain seems to lack the Crabtree effect and respires even at high external glucose levels. It also respires at a rate 4.5 times higher than the reference stain. Another alternative could be to alter the redox balance by using only the NAD-dependent glutamate dehydrogenase (GDH2), as suggested by Larsson et al..
All of these strategies could increase the ethylene yield significantly, however the theoretical yield is still quite low. Larsson et al. reported it to be 2 mol ethylene / 100 mol glucose in their study. The economic feasibility of this system is hence questionable, at least as long as the oil price remains low.
In summary, this study investigated the effects of several different cultivation factors on ethylene formation in S. cerevisiae expressing the EFE in continuous cultures. Our main finding is that oxygen availability is crucial for ethylene production. This can be coupled to two reasons: firstly, most likely EFE has a low affinity for oxygen and secondly, the requirement of oxygen for oxidation of NADH in respiration. Linked to the functioning of the EFE we also suggest that elevated arginine levels will shift the reaction towards the unwanted side reaction of EFE (reaction 3), which suggests that a process and strain optimized for ethylene production via the EFE should strive to keep intracellular arginine levels low. The ability to consume NADH seems to be a crucial requirement for optimal ethylene production, both from the previous metabolic modeling and from the data presented in this paper. An ethylene production fermentation process and strain should therefore address this issue, e.g. by ensuring a high degree of respiration and/or by changing the co-factor utilization of Gdh.
Materials and methods
Yeast strain construction
The plasmid construct (pYX212-EFE) in which the open reading frame of the EFE of Pseudomonas syringae pv. phaseolicola is cloned after the TPI1 promoter was introduced into the S. cerevisiae strain CEN.PK 113-5D (MATa MAL2-8c SUC2 ura3-52).
Aerobic chemostat cultivations were performed in carbon limited CBS minimal media as reported by Verduyn et al. with 3.5 g KH2PO4 and 0.75 g MgSO4*7H2O per liter and double amount of trace metals. As standard condition 7.5 g ammonium sulfate and 10 g glucose per liter was used. 100 ml overnight shake flask culture in the CBS media was used as inoculum for the bioreactor. The cultivations were performed in a 3 L Belach BR02 fermentor (Belach, Stockholm, Sweden) with a working volume of 2 L, operated at 30°C and with a standard aeration rate of 1.0 L min-1. The pH was kept constant at 5.0 using 1 M or 2 M NaOH.
In the initial phase the bioreactor culture was run as a batch cultivation. The switch to continuous mode was done as the batch cultivation reached late exponential phase.
Dilution rate was based on earlier measurements of pump performances. By weighing the amount of media which was pump at a certain rpm setting during a specific time period and then repeating this over a wide range of rpm settings an equation was found to calculate media flow through the pump based on the rpm. As the volume of the cultivation was known the dilution rate could be altered according to wishes. Routinely a dilution rate of 0.1 h-1 was used. A control experiment was performed to verify that the dilution rate did not affect the stripping of ethylene from the culture. The measured ethylene concentration in the off gas was identical at all conditions tested, i.e. from no dilution up to a dilution rate of 0.35 h-1.
The effect of oxygen availability on ethylene production was considered by altering the dissolved oxygen percentage. Desired oxygenation level was achieved by decreasing the stirring rate, reading the pO2 value and if necessary adjusting the rpm further. Stirrer speed was verified to not affect the stripping of ethylene from the culture over the range used. To examine the effect of nitrogen source on ethylene production, cultivations were performed with 7.5 g glutamate per liter or 3.5 g glutamate plus 3.5 g arginine per liter as alternatives to ammonium sulfate. Further, cultivations were performed at several dilution rates, from 0.033 h-1 to 0.35 h-1.
Ethylene measurement by GC-FID
Ethylene production was measured on-line by connecting the off gas to a HP 5890II GC-FID (Hewlett-Packard, USA) equipped with a HayeSep Q 80/100 porous packed column. The GC was operated isothermally at 60°C with injection temperature 70°C and detection temperature 125°C. Helium was used as carrier gas at a flow rate of 30 ml min-1 and the flame was fed with hydrogen gas. Samples were compared to an ethylene standard to determine amounts. Samples were analyzed using ChromNav (JASCO, Japan). Samples were taken automatically every second hour as standard setting.
Analysis of biomass
Cell growth (biomass) was followed by measuring OD at 610 nm. Dry weight was determined at two separate time points for each condition by the method described by Dynesen et al.. Nitrocellulose filters with pore size 0.2 μm (Sartosius Stedim Biotech, Göttingen, Germany) were pre-dried 20 min at 119 W and then weighed. A known sample volume was filtered through the pre-weighed filter papers using water suction and then washed twice with deionized water. Filters with culture sample were microwaved 20 min at 119 W, stored in a desiccator for a minimum of 24 hours before being weighed and dry weight determined.
Analysis of extracellular metabolites
Extracellular samples were collected by extracting samples from the cultivation and directly centrifuging them at 13 000 x g for 2 min. Supernatants were stored at -20°C until measured. Glucose, ethanol, glycerol, acetate, pyruvate, succinate and 2-oxoglutarate were measured using HPLC. Two separate samples were taken, with a time interval, for each condition. Ethanol was not corrected for evaporation.
The research leading to these results has received funding from the European Community's Seventh Framework Programme (FP7/ 2007–2013) under the grant agreement n°FP7-241566 – BIOCORE.
- British Petroleum: Statistical Review of World Energy June 2012. 2012, London, UK: BP GlobalGoogle Scholar
- Koottungal L: International survey of ethylene from steam crackers - 2012. Oil & Gas Journal. 2012, 110: 85-93.Google Scholar
- Mann P, Walsh PR, Mackey PJ, Jaising V, Hwang HH, Chen JC, Maheshwari M: Dan CA. 2010, Preparing for a supercycle. In Morgan Stanley Blue Paper: PetrochemicalsGoogle Scholar
- True WR: Global ethylene capacity continues advance in 2011. Oil & Gas Journal. 2012, 110: 78-84.Google Scholar
- Matar S, Hatch LF: Chemistry of Petrochemical Processes. 2001, Houston, Texas: Gulf Professional Publishing, 2Google Scholar
- Fukuda H, Ogawa T, Tanase S: Ethylene production by microorganisms. Adv Microb Physiol. 1993, 35: 275-306.View ArticleGoogle Scholar
- Nagahama K, Ogawa T, Fujii T, Fukuda H: Classification of ethylene-producing bacteria in terms of biosynthetic pathways to ethylene. J Ferment Bioeng. 1992, 73: 1-5. 10.1016/0922-338X(92)90221-F.View ArticleGoogle Scholar
- Nielsen J, Larsson C, van Maris A, Pronk J: Metabolic engineering of yeast for production of fuels and chemicals. Curr Opin Biotechnol. 2013, 24: 398-404. 10.1016/j.copbio.2013.03.023View ArticleGoogle Scholar
- Zhang J, Babtie A, Stephanopoulos G: Metabolic engineering: enabling technology of a bio-based economy. Current Opinion in Chemical Engineering. 2012, 1: 355-362. 10.1016/j.coche.2012.09.003.View ArticleGoogle Scholar
- Pirkov I, Albers E, Norbeck J, Larsson C: Ethylene production by metabolic engineering of the yeast Saccharomyces cerevisiae. Metab Eng. 2008, 10: 276-280. 10.1016/j.ymben.2008.06.006View ArticleGoogle Scholar
- Tao L, Dong HJ, Chen X, Chen SF, Wang TH: Expression of ethylene-forming enzyme (EFE) of Pseudomonas syringae pv. glycinea in Trichoderma viride. Appl Microbiol Biotechnol. 2008, 80: 573-578. 10.1007/s00253-008-1562-7View ArticleGoogle Scholar
- Wang JP, Wu LX, Xu F, Lv J, Jin HJ, Chen SF: Metabolic engineering for ethylene production by inserting the ethylene-forming enzyme gene (efe) at the 16S rDNA sites of Pseudomonas putida KT2440. Bioresour Technol. 2010, 101: 6404-6409. 10.1016/j.biortech.2010.03.030View ArticleGoogle Scholar
- Weingart H, Volksch B, Ullrich MS: Comparison of ethylene production by Pseudomonas syringae and Ralstonia solanacearum. Phytopathology. 1999, 89: 360-365. 10.1094/PHYTO.19184.108.40.2060View ArticleGoogle Scholar
- Ishihara K, Matsuoka M, Inoue Y, Tanase S, Ogawa T, Fukuda H: Overexpression and in-Vitro Reconstitution of the Ethylene-Forming Enzyme from Pseudomonas-Syringae. J Ferment Bioeng. 1995, 79: 205-211. 10.1016/0922-338X(95)90604-X.View ArticleGoogle Scholar
- Sakai M, Ogawa T, Matsuoka M, Fukuda H: Photosynthetic conversion of carbon dioxide to ethylene by the recombinant cyanobacterium, Synechococcus sp. PCC 7942, which harbors a gene for the ethylene-forming enzyme of Pseudomonas syringae. J Ferment Bioeng. 1997, 84: 434-443. 10.1016/S0922-338X(97)82004-1.View ArticleGoogle Scholar
- Ungerer J, Tao L, Davis M, Ghirardi M, Maness PC, Yu JP: Sustained photosynthetic conversion of CO2 to ethylene in recombinant cyanobacterium Synechocystis 6803. Energy & Environmental Science. 2012, 5: 8998-9006. 10.1039/c2ee22555gView ArticleGoogle Scholar
- Chen X, Liang Y, Hua J, Tao L, Qin WS, Chen SF: Overexpression of bacterial ethylene-forming enzyme gene in Trichoderma reesei enhanced the production of ethylene. Int J Biol Sci. 2010, 6: 96-106.View ArticleGoogle Scholar
- Larsson C, Snoep JL, Norbeck J, Albers E: Flux balance analysis for ethylene formation in genetically engineered Saccharomyces cerevisiae. Iet Systems Biology. 2011, 5: 245-251. 10.1049/iet-syb.2010.0027View ArticleGoogle Scholar
- Hong KK, Nielsen J: Metabolic engineering of Saccharomyces cerevisiae: a key cell factory platform for future biorefineries. Cell Mol Life Sci. 2012, 69: 2671-2690. 10.1007/s00018-012-0945-1View ArticleGoogle Scholar
- Fukuda H, Ogawa T, Tazaki M, Nagahama K, Fujii T, Tanase S, Morino Y: 2 reactions are simultaneously catalyzed by a single enzyme - the arginine-dependent simultaneous formation of 2 products, ethylene and succinate, from 2-oxoglutarate by an enzyme from pseudomonas-syringae. Biochem Biophys Res Commun. 1992, 188: 483-489. 10.1016/0006-291X(92)91081-ZView ArticleGoogle Scholar
- Boer VM, Crutchfield CA, Bradley PH, Botstein D, Rabinowitz JD: Growth-limiting intracellular metabolites in yeast growing under diverse nutrient limitations. Mol Biol Cell. 2010, 21: 198-211. 10.1091/mbc.E09-07-0597View ArticleGoogle Scholar
- Nagahama K, Ogawa T, Fujii T, Tazaki M, Tanase S, Morino Y, Fukuda H: Purification and properties of an ethylene-forming enzyme from pseudomonas-syringae Pv phaseolicola-Pk2. J Gen Microbiol. 1991, 137: 2281-2286. 10.1099/00221287-137-10-2281View ArticleGoogle Scholar
- Hahm DH, Kwak MY, Bae M, Rhee JS: Effects of dissolved-oxygen tension on microbial ethylene production in continuous culture. Biosci Biotechnol Biochem. 1992, 56: 1146-1147. 10.1271/bbb.56.1146.View ArticleGoogle Scholar
- Dejong L, Kemp A: Stoicheiometry and kinetics of the Prolyl 4-hydroxylase partial reaction. Biochimica Et Biophysica Acta. 1984, 787: 105-111. 10.1016/0167-4838(84)90113-4View ArticleGoogle Scholar
- Weusthuis RA, Visser W, Pronk JT, Scheffers WA, Van Dijken JP: Effects of oxygen limitation on sugar metabolism in yeasts - a continuous-culture study of the Kluyver effect. Microbiology-Uk. 1994, 140: 703-715. 10.1099/00221287-140-4-703.View ArticleGoogle Scholar
- Furukawa K, Heinzle E, Dunn IJ: Influence of oxygen on the growth of saccharomyces-cerevisiae in continuous culture. Biotechnol Bioeng. 1983, 25: 2293-2317. 10.1002/bit.260251003View ArticleGoogle Scholar
- Fraenkel DG: Yeast intermediary metabolism. 2011, Cold Spring Harbor, N.Y.: Cold Spring Harbor Laboratory PressGoogle Scholar
- Regenberg B, Grotkjaer T, Winther O, Fausboll A, Akesson M, Bro C, Hansen LK, Brunak S, Nielsen J: Growth-rate regulated genes have profound impact on interpretation of transcriptome profiling in Saccharomyces cerevisiae. Genome Biol. 2006, 7: R107- 10.1186/gb-2006-7-11-r107View ArticleGoogle Scholar
- Frick O, Wittmann C: Characterization of the metabolic shift between oxidative and fermentative growth in Saccharomyces cerevisiae by comparative C-13 flux analysis. Microb Cell Fact. 2005, 4: 30- 10.1186/1475-2859-4-30View ArticleGoogle Scholar
- Van Hoek P, Van Dijken JP, Pronk JT: Effect of specific growth rate on fermentative capacity of baker's yeast. Appl Environ Microbiol. 1998, 64: 4226-4233.Google Scholar
- Verduyn C, Postma E, Scheffers WA, Van Dijken JP: Effect of benzoic-acid on metabolic fluxes in yeasts - a continuous-culture study on the regulation of respiration and alcoholic fermentation. Yeast. 1992, 8: 501-517. 10.1002/yea.320080703View ArticleGoogle Scholar
- Larsson C, Nilsson A, Blomberg A, Gustafsson L: Glycolytic flux is conditionally correlated with ATP concentration in Saccharomyces cerevisiae: a chemostat study under carbon- or nitrogen-limiting conditions. J Bacteriol. 1997, 179: 7243-7250.Google Scholar
- Palmieri F, Klingenberg M: Inhibition of respiration under control of azide uptake by mitochondria. Eur J Biochem. 1967, 1: 439- 10.1111/j.1432-1033.1967.tb00093.xView ArticleGoogle Scholar
- Otterstedt K, Larsson C, Bill RM, Stahlberg A, Boles E, Hohmann S, Gustafsson L: Switching the mode of metabolism in the yeast Saccharomyces cerevisiae. Embo Reports. 2004, 5: 532-537. 10.1038/sj.embor.7400132View ArticleGoogle Scholar
- Dynesen J, Smits HP, Olsson L, Nielsen J: Carbon catabolite repression of invertase during batch cultivations of Saccharomyces cerevisiae: the role of glucose, fructose, and mannose. Appl Microbiol Biotechnol. 1998, 50: 579-582. 10.1007/s002530051338View ArticleGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.