Oregano (Carvacrol) in 60 Seconds
4 Science-Based Health Benefits of Oregano
IV:XX COVID-19 - Antiviral and AntiMalarial Terpenes in Cannabis
Terpenes: The Most Common Language in the World
http://europepmc.org/article/PMC/3821880
http://ijml.ssu.ac.ir/browse.php?a_id=196&sid=1&slc_lang=en&html=1
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
URL: http://ijml.ssu.ac.ir/article-1-196-en.html
Herpes viruses are a large family of DNA viruses with an icosahedral capsid and genomes encoding 100-200 genes. These viruses have similar morphology and replication and can make a latent and recurrent infection. Herpes simplex viruses are normal human pathogens with benign lesions but sometimes cause lethal diseases. Herpes simplex virus type 1 (HSV-1) is a member of this family which has a short growing cycle and the latency in neurons [1]. After initial infection, these viruses tend to remain dormant in sensory ganglia, so that if the patient is placed under adverse conditions such as stress, sun exposure, radiation, ultraviolet, pressure, fever, hormonal factors and so on, or the immune system is suppressed or weakened, latent virus becomes active. Primary infection occurs due to skin, mucous membranes and eye contact with infected secretions. The virus further causes wounds around the mouth, herpes labialis, scorpion disease, corneal inflamm-ation and encephalitis [1, 2].
Viral replication begins by binding virus to specific receptors. After penetration and removing the coverage, the genome is released and the expression of Immediate early, early and late last is done. Finally virus is accumulated and released, which can be accompanied by cell lysis. Two of the most important genes involved in virus replication process are early and late genes (UL52 and UL27, respectively) [3]. One of the drugs used in the treatment of injuries resulting from this virus is acyclovir. It is activated by viral thymidine which controls viral DNA polymerase and prevents viral replication in infected cells. Most chemical treatments against herpes simplex infections, which target viral proteins and involve in DNA synthesis, are almost successful in the treatment of infection caused by the virus. Nowadays, due to the mutant viruses without this enzyme, resistance to it especially in immuno-compromised people is increasing [2]. Therefore, the development of new drug compounds in order to treat diseases caused by herpes simplex virus is essential. Medicinal plants are appropriate choices for antiviral effect due to their low side effects on human. Carvacrol (2-methyl-5-1-methyl-ethyl phenol) is a phenol monoterpenoid that has shown by the previous studies to have a broad antimicrobial activity on pathogenic fungi, yeast, viruses, and bacteria [4]. Studies have indicated that monoterpenoid compounds such as carvacrol, in addition to savoring antimicrobial and anti-fungal activity are effective in anti-tumor and anti-cancer conditions [1, 5]. The antiviral effects on carvacrol has been tested on other viruses such as rotavirus and human respiratory syncytial virus [3]. In this study we investigated antiviral activity of this compound on HSV-1 and mechanism of the action of virus and its possible impact on the various stages of replication.
Materials and Methods
Carvacrol was prepared from department of physiology, in Shahid Sadoughi University of Medical Science, in Yazd, Iran. Stoke concentration was 1 mg/ml which was dissolved in dimethyl sulfoxide (DMSO) and for all experiments final concentration of DMSO was below 1% with no effect on virus and cells. Acyclovir was prepared from Amin Pharmaceutical Company in Iran and was dissolved in distilled water to make a stock concentration of 100 µM [6]. This study was approved by Ethics Committee of Shahid Sadoughi University of Medical Sciences, Yazd, Iran.
Provision and maintenance of cell lines
Vero cells (National Center of Genetic and Biological Resources, Iran) that were prepared under the supervision of Academic Center for Education, Culture and Research (ACECR) University, were grown in monolayer culture with Dulbecco´s Modified Eagle´s Medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 100 U/ml penicillin and 100 µg/ml streptomycin (all Gibco, Karlsruhe, Germany) and was kept in incubator with 5% CO2 at 37°C [6, 7].
Preparation and virus replication of HSV-1
HSV-1 strain KOS (Tarbiat Modares University, Iran) was used for experiments. HSV-1 stock cultures were prepared from supernat-ants of infected cells and were stored at -70°C. Infectivity titers were determined by a tissue culture infectivity dose (TCID50) method [8].
Cytotoxicity test
In order to achieve a concentration of drug which was nontoxic on vero cells, [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide] (MTT) test was used. In this test, cells were seeded into the 96 well plates. Plates were kept in incubator at 37°C with 5% CO2 for 24 hrs. Then, the cells were washed with preheated phosphate buffered saline (PBS), and serial dilution of carvacrol from 0.1% to 0.00001% was added to vero cells. After 48 hrs of incubation at 37°C, MTT dye uptake was determined by measuring the optical density at 570 nm in a spectro-photometer (Biotek Instrument Model: Box 998, United States). Wells containing medium with 1% DMSO were used as control [9].
Determine the viral titer by TCID50 method
To determine the appropriate concentration of the virus that makes pathological change in 50% of the cells, TCID50 method was used. Virus infectivity was quantified by estimating the 50% TCID50 using standard cell culture procedures. Briefly, when cells reached 80% confluency 96-well microtitre plates, six replicates were infected with 1.8 ml DMEM combination without serum as a diluents and 0.2 mol virus, and then were incubated for 1 hr at 37°C in a humidified atmosphere of 5% CO2 until viruses were absorbed into the cell. Afterwards, the media was removed and DMEM containing 2% FBS was added into 96-well plates and incubated for 48 hrs at 37°C. Cytophatic effect was checked by an inverted microscope (ACCU-SCOPE, United States) and infectivity titers were expressed as TCID50/ml based on the Karber formula. The virus infectivity was then compared and analyzed to determine the optimal sample formulation under the different conditions of preparation and storage [10, 11].
Antiviral activity
The antiviral activity of carvacrol was evaluated with TCID50 assay. Different concentration of the drug was incubated with virus for 1 hr at 25°C. Then the resulting compound was inoculated to the cells for 1 hr at 37°C, and then the cells were subjected to TCID50 assay. The 50% inhibitory concentration (IC50) of the compound was evaluated from dose response curves [12-14].
HSV-1 incubation with carvacrol prior to infect vero cells with virus
In this experiment, the ability of drugs to bind to the virus and change their ability to infect cells was tested. For this purpose, the desired concentration of HSV-1 virus with non-toxic concentrations of carvacrol was incubated for 1 hr at room temperature and then was added to the cells and removed from the cells and media containing 2% FBS was added to the cells. The results were recorded after 24 to 48 hrs by TCID50 assay [15]. Beta pinene was used as a monoterpene compound control which had been prepared from Carl Roth (Karlsruhe- Germany). This compound is solved in ethanol; during testing we noted that final ethanol concentration should not exceed 1%.
Treatment of vero cell before infecting cells with HSV-1
Drugs were examined for their ability to bind to cell receptors and to prevent the virus. For this purpose cells and drug were incubated for 1 hr at 37°C. Then the cells were washed and HSV-1 with defined concentration of carvacrol were added to the cells and were incubated for 1 hr at 37°C. Afterwards, the medium were removed, and DMEM medium containing 2% FBS were added and the cells were checked after two days.
Vero cell incubation with carvacrol after cell infection with HSV1
The ability of combination in affecting virus infection in intercellular were tested. After the penetration of the virus to infected cells, drug was added to cells. Wells containing medium with 1% DMSO but no compound were also used as control. After two days, cells were subjected to TCID50 assay [16].
Attachment assay
To determine antiviral effect of carvcrol on attachment of virus, vero cell monolayers were grown in 24-well culture plates and then prechilled at 4°C for 1 hr. After aspirating the medium, cells were infected with virus in the absence of the maximum noncytotoxic carvacrol concentrations. Then, cells were kept at 4°C for another 3 hrs. After removing the medium, cells were washed with PBS [17-19]. Finally, the effect of carvacrol on attachement was checked with TCID50 assay. Melissa officinalis with a stock concentration of 1.43 mg/ml was used as a positive control [16].
The evaluation of HSV-1 penetration to cells
To determine the effect of carvacrol on viral penetration, cells were prechilled at 4°C for 1 hr followed by infection with virus for 2 hrs at 4°C. Carvacrol was added for another 30 min. at 4°C. To allow viral penetration, the temperature of incubator was changed to 37°C. After 30 min. at 37°C, cells were treated with citrate buffer (135 mM NaCl, 10 nM KCl, 40 mM sodium citrate, pH 3) to stop penetration and to inactivate attached, unpenetrated virions. After removing the buffer, the cell was overlaid with medium and then the effect of carvacrol on penetration was checked with TCID50 assay [18].
Real Time polymerase chain reaction (RT-PCR)
The level of gene expression of late and early genes were determined by RT-PCR. For this purpose, vero cells were infected with HSV-1 in 24 well plates at 37°C for 1 hr, then cells were overlaid with medium and carvarol. Acyclovir was used as a control. After 48 hrs, the cellular RNA was extracted using a kit from RIBO-PREP Company (Moscow, Russia). Then, reverse transcription of RNA was performed by cDNA synthesis kit (Fermantas, USA), yielding cDNA. In the presence of specific primers, the cDNA was used for RT-PCR using SYBR green. The primer pairs for UL52 were forward (5′GACCGACGGGTGCGTTATT3′) and reverse (5′GAAGGAGTCGCCATTTAGCC3′); for UL27 were forward (5′GCCTTCTT CGCCTTTCGC3′) and reverse (5′CGCTCGTGCCCTTCTTCTT3′) [20]. Glyceral-dehyde 3-phosphate dehydro-genase (GAPDH) was used as the reference gene which primer pairs were GAPDH forward (5′CCCACTCCTCCACCTTTGAC3′) and GAPDH reverse (5′TCTTCCTCTT GTGCTCTTGC3′). Comparison of gene expressions was performed by the delta delta Ct method.
Results
The results of the toxicity of carvacrol on vero cells using MTT
The effect of different concentrations of the drug on vero cells MTT assay processed for 48 hrs led to, a desired result (TC50 was 0.001%).
Determination of 50% inhibitory concentra-tion of carvacrol against virus
In order to determine inhibitory concentration of carvacrol against virus, non-toxic drug with different concentrations of carvacrol was exposed to the virus and then was inoculated into the cells. The concentration range tested for carvacrol was up 0.00001% - 0.00075%, which was obtained by MTT assay. 50% of inhibitory concentration of carvacrol against virus was 0.0002% (Table 1).
Results of carvacrol antiviral mechanism against HSV-1
The results of virus titer by TCID50/ml was 104.5. Results show that carvacrol decreases herpes up to 70%. There was no reduction when drug enters the cell before entering the virus. The result showed that Carvacrol has no effect on replication. Carvacrol has no effect in attachment. Based upon data, carvacrol has no effect on penetration (Fig. 1).
The results of early and late virus gene expression of HSV-1
RT-PCR result revealed that this drug had no role in the early and late gene expression (Fig. 2).
Table 1. The results of the drug selectivity index | Drug | Max noncytotoxic concentration | TC50 | IC50 | SI (TC50/IC50)a |
| Carvacrol | 0.00075% | 0.001% | 0.0002% | 5 |
| Acyclovir | 100 μm | ≥ 100 μm | 1.8 μm | ≥ 56 |
TC50= 50% toxic concentration; IC50= 50% inhibitory concentration
Discussion
Herpes virus can cause lifelong infection. This infection is latent and is associated with reactivation. Many of the known human herpes viruses infect most of the world's population. HSV-1 leads to different clinical symptoms from hepetic labialis to an acute encephalitis in humans [1, 2]. Drug for the treatment of diseases caused by this virus is acyclovir. However, due to prolonged use of the drug, resistant strains of the drug have been identified. Acyclovir resistant strains can cause severe clinical consequences in patients [2].
The use of medicinal plants to treat and prevent various diseases has been widely growing. Recently, due to drug resistance and its side effects, use of drugs of plant origin for the treatment of diseases has increased. In this study, we tried to investigate the effect of carvacrol combination on HSV-1. Many other studies have been conducted to determine the effects of plant extracts on HSV-1. A study by Khanavi and colleagues in 2009 presented the chemical components of essential oils of thyme and oregano. It was found that essential oil of thyme twigs air contains 24 compounds including: thymol (38%), carvacrol (34.96%), p-cymene (7.17%) and beta-caryophyllene (2.71%) [21]. In another study by Cinati and colleagues it was shown that one of the constituents of licorice is carvacrol [22].
Based on this study and other studies by Mardani et al. [23], Farahani et al. [24], Sabouri Ghanad et al. [25] and Monouri et al. [26] on plant extract, our study was determined to check antiviral activity of carvacrol on vero cells. Carvacrol is a compound with the scientific name Methyl ethyl 2-methyl phenol 5-1 which has an anti-bacteria and anti-fungi effects. Magi et al. in 2015 showed the antimicrobial effect of carvacrol on Streptococcus pyogenes [27]. Maximum nontoxic concentration of this compound determined with MTT assay was 0.0001.
The antiviral effect of the drug was tested and IC50 was 0.0002. With TC50/IC50, selectivity index was calculated which was 5. According to the research that was conducted by Amoros et al. in 1992, the drugs with selectivity index higher than 4 are considered preferable [28]. Astani et al. worked on some essential oils and monoterpenoids and some of their selectivity indexes were as follows: Thyme oil: 6.4, Thymol: 2.8, Citral: 12.9 [29].
According to this study, carvacrol which belonged to monoterpenoids and one of the compounds of plants essential oils had similar selectivity index. In a study conducted by Mardani et al. in Iran in 2012, the antiviral effect of Shirazi thyme essential oils, was determined. Carvacrol was one of the compounds that its antiviral effect against HSV-1 was considered. In this pilot study, cytotoxicity of essential oils in different concentrations was performed on vero cells. The result showed that the concentration which destroyed 50% of vero cells was 0.0676, and the concentration which inhibited detection of viral plaque was 0.0059. Also tested essential oils in 0.01, 0.02 could prevent the virus completely [23]. Based on the previous studies, thyme essential oils can prevent the virus completely. In this study, carvacrol, as a part of thyme essential oils, reduced the virus up to 70%, so it showed that another compound of thyme essential oils has antiviral effect. The concentration used was 0.01 and 0.02 %. However, in this study the concentration of 0.00075% was inhibited by 70% that is a good result for carvacrol compound.
In another study which was performed by Farahani in 2012, the antiviral effect of thyme on HSV-1 was tested in vitro. The antiviral effect of this plant was considered by preventive effects of cytopatic virus. Thyme essential oils prevent herpes virus proliferation and in non-toxic concentration shows antiviral effect on HSV-1 [24].
Another study by Pilao and his colleagues on Mentha pulegium essential oils showed that carvacrol is the main compound thus its antiviral effect on human and animals was investigated. They concluded that these essential oils are able to inhibit different human and animal virus such as Rota virus, bovine diarrhea virus and respiratory syncytial virus in vitro [15]. In the next phase, drug mechanism has been considered. This experiment was performed in three methods: incubation of HSV-1 with carvacrol before infection with vero cells with the virus (Pretreatment of virus), treatment of vero cells with carvacrol before contaminating cells with HSV-1 (Pretreatment of cell), and incubation of vero cells with carvacrol after infecting the cells with HSV-1 (replication). Carvacrol shows 70% decrease in pretreatment of virus. In another study carried out by Monavari and colleagues on thyme essential oils in 2013, the antiviral effect of this extract against HSV-1 before infection cells with the virus was probed. In this article, Melissa and B-pinene were used as positive control which both showed 100% decrease [26].
In this study, in pretreatment of the cell, only Mellissa showed 60% decrease. In replication, carvacrol has no decrease because it cannot enter the cell while acyclovir, as a positive control, has 100% decrease which is due to its mechanism that affects on replication. After this stage, antiviral effect of carvacrol on penetration and attachment modes was checked which showed no inhibitory effect on virus. In final stage, primary and delay gene was checked by RT-PCR method, and it resulted in that carvacrol shows no change in none of them although acyclovir showed decrease for UL52 and UL27 40% and 20%, respectively. It means that it plays important role for gene expression decrease. We resulted that the use of plant drugs is more effective and has no side effects; carvacrol compound showed ant-viral effects in pretreatment process.
This study suggests that compounds such as Carvacrol and beta-pinene are monoterpenes and the synergism between them can be considered as a next step. We reached a good conclusion on low toxicity and anti-herpes virus by in vitro study. In vivo studies can be studied on animal laboratory.
Conflict of Interest
The authors declare that they have no competing interests.
Acknowledgment
There is no acknowledgment to declare.
References
- Brooks GF, Carroll KC, Butel JS, Morse SA. Medical microbiology. 24th ed, US: McGrawHill; 2007.
- Greenberg MS, Glick M. Burket`s oral medicine. 11th ed, USA: BC Decker Inc; 2008. pp. 42-46.
- Pilau MR, Hartz Alves S, Weiblen R, Arenhart S, Cueto AP, Lovato LT. Antiviral activity of the Lippia graveolens (Mexican oregano) essential oil and its main compound carvacrol against human and animal viruses. Brazil J Microbiol. 2011; 42(4): 1616-624.
- Woo SB, Greenberg MS. Ulcerative, vesicular and bullous lesions. In: Greenberg MS, Glick M, Ship JA, editors. Burket's Oral Medicine. 10th ed, New Delhi: BC Decker; 2007. pp. 41-76.
- Freeman DL. Harrison's principles of internal medicine. JAMA. 2001; 286(8): 971-72.
- Astani A, Heidary NM, Schnitzler P. Attachment and penetration of acyclovir resistant herpes simplex virus are inhibited by melissa officinalis extract. Phytother Res. 2014; 28(10): 1547-552.
- Schnitzler P, Nolkemper S, Stintzing FC, Reichling J. Comparative in vitro study on the anti-herpetic effect of phytochemically characterized aqueous and ethanolic extracts of Salvia officinalis grown at two different locations. Phytomedicine 2008; 15(1-2): 62-70.
- Kukhanova MK, Korovina AN, Kochetkov SN. Human herpes simplex virus: lifecycle and development of inhibitors. Biochem. 2014; 79(13): 1635-652.
- De Jalon EG, Blanco-Prieto MJ, Yqartua P, Santoyo S. Increased efficacy of acyclovir-loaded microparticles against herpes simplex virus type 1 in cell culture. Eur J Pharm Biopharm. 2003; 56(2): 183-87.
- Corey L, Adams H, Brown A, Holmes K. Genital herpes simplex virus infections, clinical manifestation, course and complications. Ann Intern Med. 2005; 98(6): 958-72.
- Arduino PG, Porter SR. Herpes simplex virus type 1 infection: overview on relevant clinic-pathological features. J Oral Pathol Med. 2008; 37(2): 107-21.
- Pastorino B, Baronti C, Gould EA, Charrel RN, Lamballerie X. Effect of chemical stabilizers on the thermostability and infectivity of a representative panel of freeze dried viruses. PloS one. 2015; 10(4): e0118963.
- Gilbert C, Bestman SJ, Boivin G. Resistance of herpes viruses to antiviral drug: clinical impacts and molecular mechanisms. Drug Resistance Updates 2002; 5(2): 88-90.
- Bacon TH, Levin MJ, Leary JJ, Sarisky RT, Sutton D. Herpes simplex virus resistance to acyclovir and penciclovir after two decades of antiviral therapy, Clinical microbiology reviews. 2003; 16(1): 114-28.
- Pilau MR, Alves SH, Weiblen R, Arenhart S, Cueto AP, Lovato LT. Antiviral activity of the Lippia graveolens (Mexican oregano) essential oil and its main compound carvacrol against human and animal viruses. Braz J Microbiol. 2011. 42(4): 1616-624.
- Astani A, Schnitzler P. Antiviral activity of monoterpenes beta-pinene and limonene against herpes simplex virus in vitro. Iran J Microbiol. 2014; 3(6): 150-55.
- Cheng HY, Lin TC, Yang CM, Wang KC, Lin LT, Lin CC. Putranjivain A from Euphorbia jolkini inhibits both virus entry and late stage replication of herpes simplex virus type 2 in vitro. J Antimicrob Chemother. 2004; 53(4): 577-83.
- Gescher K, Hensel A, Hafezi W, Derksen A, Kühn J. Oligomeric proanthocyanidins from Rumex acetosa L. inhibit the attachment of herpes simplex virus type-1. Antiviral Res. 2011; 89(1): 9-18.
- Astani A, Reichling J, Schnitzler P. Melissa officinalis extract inhibits attachment of herpes simplex virus in vitro. Chemother. 2012; 58(1): 70-77.
- Yangfei X, Ying P, Chang Q, Zhicai L, Zhe R, Ke Y, et al. In vitro anti herpes simplex virus activity of 1,2,4,6 Tetra O galloyl - β - D glucose from Phyllanthus emblica L. Euphorbiaceae. 2011; 25(7): 975-82.
- Khanavi M, Norouzi M, Tabatabaee H, Noudeh A, Safavi S, Shafiee A. Chemical compositions and antiviral effects of the essential oil of zataria multiflora boiss and origanum majorana L. JMP. 2010; 1(33):128-37.
- Cinatl J, Morgenstern G, Bauer G, Chndra P, Rabenau H, Doerr W. Glycyrrhizin, an active component of liquorice roots, and replication of SAR-associatde coronavirus. Lancet 2003; 361: 2045-2046.
- Mardani M, Motamedifar M, Hoseinipour R. A study of the antiviral effect of the essential oil of zataria multiflora boiss on herpes simplex Type 1 in vero cell culture. J Dent Shiraz Univ Med Sci. 2012; 13: s414-s420.
- Farahani M. Antiviral effect assay of thymus kotschyanus and nerium oleander on HSV-1 in vitro. JSSU. 2013; 21(2): 189-97.
- Sabouri Ghanad AM, Safiallahy S, Faradmal J. The evaluation of methanolic extract of glycyrrhiza glabra effect on the replication of herpes simplex virus Type 1 in vero cell line. Avicenna Journal of Clinical Medicine 2014; 20(4): 325-32.
- Monavari SHR, Shamsi Shahrabadi M, Mortazkar P. the study of antiviral effects of glycyrrihza glabra extract on HSV. J Med Plants. 2008; 4(28): 81-86.
- Magi G, Marini E., Facinelli B. Antimicrobial activity of essential oils and carvacrol, and synergy of carvacrol and erythromycin, against clinical, erythromycin-resistant group A streptococci. Front Microbiol. 2015; 6: 165.
- Amoros M, Simoes CMO, Girre L, Sauvager F, Cormier M. Synergistic effect of flavones and flavonols against herpes simplex virus type 1 in cell culture. Comparison with the antiviral activity of propolis. J Nat Prod. 1992; 55: 1732-740.
- Astani A, Reichling J, Schnitzler P. Comparative study on the antiviral activity of selected monoterpenes derived from essential oils. Phytother Res. 2010; 24(5): 673-79.
https://pubs.acs.org/doi/pdf/10.1021/jp108675b
Inclusion of Terpenoid Plant Extracts in Lipid Bilayers Investigated by Molecular Dynamics Simulations†
- Sarah Witzke ,
- Lars Duelund ,
- Jacob Kongsted ,
- Michael Petersen ,
- Ole G. Mouritsen , and
- Himanshu Khandelia
Abstract
The plant Perilla frutescens is widely employed in Asian medicine. The active components of Perilla include cyclic terpenes, which have a diverse range of antimicrobial, anticancer, sedative, and anti-inflammatory properties, hinting at a membrane-mediated mechanism of action. We have used molecular dynamics (MD) simulations and isothermal titration calorimetry (ITC) to investigate the interaction of four terpenes with model lipid bilayers. The ITC and MD data are mostly in accordance. The terpenes partition into membranes, pack along the lipid tails, and alter bilayer structure and dynamics. Three of the four molecules could cross the bilayer. The carboxylate-group-containing terpene modifies headgroup repulsion and increases the area per lipid by more than 10%, in a manner reminiscent of membrane-thinning peptides and solvents such as DMSO. Our results support the possibility that at least some medicinal properties of volatile Perilla extracts might arise from interactions with the lipid bilayer component of biological membranes.
https://www.osti.gov/pages/servlets/purl/1484588
Membrane Permeability of Terpenoids Explored with Molecular Simulation Josh V. Vermaas,† Gayle J. Bentley,‡ Gregg T. Beckham,∗,‡ and Michael F. Crowley∗,† †Biosciences Center, National Renewable Energy Laboratory, Golden, CO 80401 ‡National Bioenergy Center, National Renewable Energy Laboratory, Golden, CO 80401 E-mail: gregg.beckham@nrel.gov; michael.crowley@nrel.gov 1 Pursuant to the DOE Public Access Plan, this document represents the authors' peer-reviewed, accepted manuscript. The published version of the article is available from the relevant publisher. Abstract Terpenoids constitute a class of compounds with remarkable potential for pharmaceutical, fragrance, specialty chemical, and biofuel applications. However, their industrial production is limited by their rarity within their native plant hosts, creating considerable interest in microbial hosts capable of manufacturing terpenoids. To reduce production costs, non-destructive product recovery from these microbial hosts is preferred, and is achievable using a hydrophobic organic overlay. Our prior research has indicated that oxidized fatty acyl products may permeate faster through host membranes, increasing overall biorefinery productivity. To test this hypothesis for terpenoids, we computed membrane permeabilities of conventional terpenoid target products (e.g. limonene, bisabolene, farnesene), and related oxidized compounds through molecular dynamics simulations. These simulations indicate that terpenoid product permeabilities from cytosol to overlay are oxidation independent, as increases in membrane extraction efficiency due to product oxidation are proportionally offset by decreases in the membrane crossing rate if the membrane and organic phase are in close contact. However, if aqueous extraction is required, oxidation will accelerate the slow product extraction from the membrane. Experimental toxicity assays performed indicated that most terpenoids tested were tolerated by microbial hosts, although exposure to oxidized terpenes often retarded microbial growth compared with conventional terpenes. Thus, terpenoid oxidation is not expected to significantly increase or decrease the extraction productivity in an industrial setting where cells are in close contact, unlike the previously studied fatty acyl products. 2 Pursuant to the DOE Public Access Plan, this document represents the authors' peer-reviewed, accepted manuscript. The published version of the article is available from the relevant publisher. Introduction Terpenes and terpenoids, naturally-occurring compounds primarily extracted from plants, display diverse functionality which allow for their use in a wide variety of chemical applications.1,2 Still, it remains difficult to rely on plant-based extraction for the large-scale production of terpenes naturally produced in low quantities from plants or animals that are difficult to cultivate, or are derived from endangered species. Compounded with these limitations is crop yield variability, further affecting the yield and types of terpenes extracted by plant-based methods.3 In response to these limitations, many research groups have successfully engineered microbes for the production of an increasingly wide range of terpenes.4–10 Additionally, engineering microorganisms for terpene production increases the space of terpene chemistries that are accessible for overproduction due to the abundance of terpene synthases that enable unique chemistries.11 From a process perspective, toxicity and ease of extraction of the terpene product are critical factors that affect the feasibility of scalable and cost effective production. Terpenes can function as antimicrobial agents to protect their natural hosts, with antibacterial activity occurring via disruption of the lipid membrane.12–14 The use of a hydrophobic organic overlay such as dodecane is often used at bench scale during microbial production of terpenes to eliminate loss of a volatile product and to reduce the immediate toxicity of the terpene product.9,10,15–17 Here, we examine the effect of terpene oxidation on cellular toxicity and extraction efficiency of the terpene or terpenoid into an organic overlay. Since terpenoid extraction faces the same challenges as extraction of other biological products, it may be informative to apply findings from other systems to terpenoids. Using molecular simulation, our recent results have indicated that hydroxyl groups maximize fatty acid extraction into a hydrophobic organic phase.18 Fatty alcohol and fatty aldehyde products were shown to efficiently cross lipid bilayers in addition to reducing the barrier to extraction of these typically lipophilic molecules. Terpenoid target molecules are generally devoid of oxygenation and are highly lipophilic, although recent work has also shown pathways towards 3 Pursuant to the DOE Public Access Plan, this document represents the authors' peer-reviewed, accepted manuscript. The published version of the article is available from the relevant publisher. creating oxidized terpenoids,19–24 and so there is a significant potential for improvement if the trend observed for fatty acyl products18 is similar for terpenoids. Bisabolene OH Bisabolol Limonene OH Terpineol OH Perillyl alcohol Pinene OH Verbenol O Verbenone HO Farnesene Farnesol Figure 1: Terpenoid compounds and their oxidized equivalents considered in this study. The colored boundaries group together equivalent chemical functionalities, which are used consistently in other visualizations throughout the text. Red highlights bisabolene and its predominant oxidized form, bisabolol, blue highlights pinene and its derivatives verbenol and verbenone, green highlights limonene with derivatives terpineol and perillyl alcohol, and violet surrounds farnesene and its oxidation product farnesol. To test whether hydroxylation of terpenoids improves the extraction rate from cell membranes to an organic overlay, in this study we have conducted molecular dynamics (MD) simulations to determine the membrane permeabilities and expected membrane extraction rates of both oxidized and non-oxidized terpenoid compounds (Fig. 1). The oxidized terpenoids studied here have biological synthetic routes, with examples of metabolic engineering to produce farnesol,19,20 bisabolol,21 verbenol,22 verbenone,23 terpineol,24 and perillyl alcohol.25 The MD findings reflect the previous trend observed for fatty acyl products, in that adding a hydroxyl group improves membrane extraction rates while retarding membrane 4 Pursuant to the DOE Public Access Plan, this document represents the authors' peer-reviewed, accepted manuscript. The published version of the article is available from the relevant publisher. crossing. Since the determined crossing and extraction rates are so similar, we predict that terpenoid oxidation does not significantly improve the theoretical extraction rate of organic phase extraction if the cell and organic phase are in close contact. However, if cells avoid organic phases through an active mechanism, as has been suggested previously based on microscopy measurements,26 our results demonstrate that oxidized terpenoids may improve extraction rates if the cell’s immediate environment is predominantly aqueous. In addition, we find through microbial growth experiments that most terpenoid products are tolerated at current experimental titers of up to 3%, although some oxidized products exhibited greater toxicity, resulting in decreased relative to their lipophilic equivalents. Thus, targeting oxidized terpenoids may be a viable path forward in optimizing microbial production of these species, depending on the environmental conditions within industrial settings. Methods Both computational and experimental assays were conducted. The molecular simulations were used to determine the permeability of these compounds. Experimental assays measure the impact of alternate terpenoids on growth. The computational methods used to assess the bilayer crossing and organic phase partitioning of terpenoids are largely derived from prior studies on fatty acyl products18 and the wider membrane permeability literature,27 which details the theory behind computing permeability coefficients separately for extraction and crossing steps. In brief, equilibrium MD simulations were performed to monitor spontaneous membrane crossing of the terpenoid compound present. These equilibrated systems formed the basis for Replica Exchange Umbrella Sampling (REUS)28 calculations used to determine local diffusivity and free energy profile for membrane crossing and extraction processes for all of the terpenoids indicated in Fig. 1. The diffusivity and free energy profiles are used to 5 Pursuant to the DOE Public Access Plan, this document represents the authors' peer-reviewed, accepted manuscript. The published version of the article is available from the relevant publisher. compute membrane permeability (P) using the inhomogeneous solubility diffusion model:29,30 Z ξu exp (ΔG (ξ) β) P −1 = dξ (1) D (ξ) ξl In this formalism, the free energy change (ΔG) and diffusivity (D) are known along the reaction coordinate (RC) ξ, which contributes to the permeability along the given RC. The details of the simulation system preparation are described in the subsequent subsections. To understand what the practical effect these compounds might have on their toxicity and growth of the host organism, which is difficult to assess computationally, growth assays were also carried out, with their methods detailed at the end of this section. Simulation System Setup Two types of simulation systems are required to accurately describe the terpenoid extraction process, one with the product in aqueous solution (Fig. 2A) representing the terpenoid adsorption to the membrane, and a second (Fig. 2B) representing terpenoid extraction from the membrane into a proximal organic phase, in this case dodecane. We consider this proximal organic phase to be a better descriptor of how dodecane interacts with cells at experimental dodecane concentrations, as scenarios where the dodecane either is sandwiched between the membrane leaflets or extracts significant lipids from the membrane would likely lead to cell death by disrupting the membrane. Both the aqueous and organic systems share a common membrane composition with the comparison study of fatty acyl product extraction to aid comparison,18 with each leaflet having a 30:14:12:9:4:1 molecular composition of phosphatidyl choline (PC), phosphatidyl ethanolamine (PE), phosphatidyl inositol (PI), ergosterol, phosphatidyl serine (PS), and phosphatidic acid (PA) headgroups and sterols and a 2:1:1 mixture of palmityl oleyl, stearyl-linoleyl and dioleyl lipids. This composition was initially chosen as a stand-in for an oleaginous yeast membrane based on the available experimental evidence for both headgroups31–33 as well as tail compositions.34,35 The membrane 6 Pursuant to the DOE Public Access Plan, this document represents the authors' peer-reviewed, accepted manuscript. The published version of the article is available from the relevant publisher. A B Figure 2: Example states near the start of simulation, detailing the relative positioning and abundance of the modeled dodecane overlay (tan) compared with the membrane (atomically- – colored spheres), ions (yellow for Na+, cyan for Cl ) and the single terpenoid product per simulation system (green). Hydrogen atoms are omitted for clarity, and the surrounding water is represented as a transparent blue surface where solution is present. (A) The overlayfree system is used as a control. (B) A high-dodecane overlay, representative of a lab-scale overlay in close proximity with the membrane. In addition to the main findings regarding membrane permeability, three of these high-dodecane overlays were considered to predict the membrane structural changes depending on the degree of bilayer hydration. The three initial spacings between the membrane and the organic phase were 5, 10, and 15 ˚A, with water to lipid ratios of approximately 12, 27, and 38 as the result. Here, we only show the 5 ˚A case. 7 Pursuant to the DOE Public Access Plan, this document represents the authors' peer-reviewed, accepted manuscript. The published version of the article is available from the relevant publisher. was assembled using CHARMM-GUI.36–38 For each terpenoid in Fig. 1, both aqueous and organic systems were initially constructed with a single terpenoid molecule added outside of the bilayer, followed by solvation and ionization using the Solvate and Autoionize plugins of VMD39 to add water and dodecane as appropriate for the system (Fig. 2). Due to the observed ordering of the lipids under conventional constant pressure simulation that are discussed later, aqueous solvation layers between the membrane and organic phase (Fig. 2B) of different thicknesses were tested, resulting in water to lipid ratios of 12, 27, and 38, depending on the degree of hydration. These three systems cover the range of lipid ratios used in previous simulation studies of lipid structural changes in response to hydration.40–42 In all instances, the total aqueous salt concentration is 150 mM NaCl, which includes the 30 ion sodium bias that results from the – net charge of the bilayer (e.g. there are 34 Na+ and 4 Cl ions in the organic system with the least water, shown in Fig. 2B). All systems have an initial total size of approximately 65 A˚×65 ˚A×100 A. ˚ Simulation Details Equilibration for each system and biased simulations used to compute membrane permeability share specific simulation parameters, which are described here. Simulations were performed with NAMD 2.12,43 using the CHARMM36 force field for lipids42 and sterols.44 The dodecane overlay and the terpenoids used parameters from the CHARMM general force field45 (CGenFF) as determined by the ParamChem webserver.46 Consistent with the standard parameterization methodology for CHARMM force fields,45 simulations were performed using a 12 ˚A nonbonded cutoff and the TIP3P water model.47 Long-range electrostatic interactions were treated using particle mesh Ewald48,49 with a 1.2 ˚A grid. A Langevin thermostat using γ=1 ps−1 maintained the system temperature50 at either 300 K (equilibration) or 310 K (biased simulation). Anisotropic pressure coupling was maintained via the Langevin piston method51,52 to a pressure of 1 atm, with periodic cell growth along the membrane normal 8 Pursuant to the DOE Public Access Plan, this document represents the authors' peer-reviewed, accepted manuscript. The published version of the article is available from the relevant publisher. axis decoupled from growth along the membrane parallel axes. In most cases, the membrane parallel axes were allowed to move freely; however, to combat observed membrane ordering, one set of REUS calculations was carried out in a fixed area ensemble, where the only dimension controlled by the barostat is along the membrane normal. Each simulation timestep was 2 fs, enabled by using the SETTLE algorithm53 to fix bond lengths to hydrogen. Equilibrium Simulation Prior to biased simulation, five independent 100 ns equilibrium simulations were performed starting from the structures shown in Fig. 2 to increase sampling of the single terpenoid such that the terpenoids within the subsequent steered MD (SMD) trajectories would not experience the same membrane environment. The 100 ns timescale permits the membrane to adopt its preferred geometry given the imposed simulation conditions and for the terpenoid positions to be thoroughly decorrelated, minimizing the bias from the initial geometry when seeding the initial positions for the REUS calculations. However, it must be emphasized that this timescale is significantly shorter than what is required for significant mixing, clustering, or leaflet exchange of lipids, which have been previously demonstrated to occur on up to the multi-microsecond timescale.54–56 Further exploration of lipid ordering within our simulated membranes was conducted through additional sets of five independent simulations of each terpenoid for 20 ns each where additional hydration is present, as described in Fig. 2B. Bilayer Crossing and Extraction Preparation For many of the terpenoid compounds shown in Fig. 1, there are comparatively few rotatable bonds, as the ring structures present constrain the conformational space accessible to each compound. Thus, for pinene, bisabolene, limonene, and their derivatives, the center of mass of the molecule accurately describes the progress along the extraction and crossing RC. For these molecules, we prepared a complete trajectory where the compound center of mass ranged from the membrane midplane to the plane 45 ˚Aaway from the bilayer center. This 9 Pursuant to the DOE Public Access Plan, this document represents the authors' peer-reviewed, accepted manuscript. The published version of the article is available from the relevant publisher. was accomplished by taking the five independent positions resulting from the final snapshot of the equilibrium trajectory, and using SMD to move the molecules to the end points of the RC in two separate simulations (“forward” along the RC and “backward”). The force A−2 constant used for this step was 7 kcal mol−1 ˚ . The total time for the “forward” and “backward” simulations was 10 ns, with the length of each simulation dependent on the distance between the initial state and the final target for the RC. For farnesene and farnesol, however, there is considerable conformational flexibility through rotation of the single bonds, which makes the center of mass an unreliable estimator of extraction progress.57,58 Similar to the approach taken when studying fatty acyl compounds,18 we split the RC into two steps, creating independent RC for bilayer adsorption/desorption and bilayer crossing. For bilayer crossing, the assumed RC, measuring the position the end of the molecule where the alcohol would be on farnesol, ranged from -25 ˚A to 25 A˚ along the membrane normal, with the zero-point located at the membrane midplane. This RC intentionally overestimates the 40 ˚A thickness of the bilayer59 so that we can be sure the free energy minimum is captured by this RC. As was done for the other compounds, “forward” and “backward” simulations were performed to take the end state from the equilibrium trajectories to the boundaries of the RC. These pulling trajectories were performed over 10 ns A−2 in total using a force constant of 2 kcal mol−1 ˚ . For extraction of farnesene and farnesol, we compose a RC based on the number of contacts between the single terpenoid compound and the membrane and, if present, the dodecane overlay. This RC is implemented as a coordination number collective variable from the colvars60 module of NAMD,43 using the same cutoffs and parameters as were used for fatty acyl products:18 XX � 4 1 − |xi − xj | /8˚A C (g1, g2) = � (2) 10 1 − |xi − xj i∈g1j∈g2 | /8˚A The coordination number is defined over the farnesene heavy atoms (g1), with g2 defined 10 Pursuant to the DOE Public Access Plan, this document represents the authors' peer-reviewed, accepted manuscript. The published version of the article is available from the relevant publisher. as either the even-numbered acyl tail carbons of the lipids, creating a lipid-farnesyl contact number (Clipid), or as a subset of the dodecane heavy atoms, creating a dodecane-farnesyl contact number (Cdodecane). For extraction into aqueous solution, we drive Clipid to 0. For extraction into an organic phase, we drive the sum of Clipid − Cdodecane to be as negative as possible. Due to greater direct access to the heavy atoms offered by the unsaturations present in farnesene and farnesol, the coordination numbers during equilibrium simulation tend to be somewhat larger than they were in the fatty acyl products (Fig. S1). Thus, the extraction RC from membrane to aqueous solution ranges from [0,650], and from [-850,650] for extraction to organic solution. To construct continuous pulling trajectories to seed the REUS calculations, each of the 5 end states from the equilibrium trajectory were pulled with “forward” and “backward” simulations with total duration of 10 ns for each replicate. The extraction pulls used a 0.01 kcal mol−1 contact−2 force constant. The five independent contiguous trajectories created by the “forward” and “backward” simulations for each compound were used to seed the initial coordinates for the REUS calculations, with each trajectory contributing an equal number of initial positions randomly distributed across each RC considered. In this way, we also sample from different places within the heterogeneous membrane environment created by the mixed bilayer, rather than simply one potential path of bilayer translocation, increasing the accuracy of the result. Replica Exchange Umbrella Sampling Parameters For each of the three distinct RC sampled (center of mass, farnesene translocation, and farnesene extraction), the REUS simulations themselves are all 20 ns long. Exchanges between alternating replica neighbors were attempted every picosecond, yielding a 2 ps overall exchange rate, consistent with reported best practices suggesting frequent exchange attempts. 61 Other parameters changed depending on the RC under study. For center of mass translocations, the [0,45] ˚A RC was divided into 60 equal umbrellas A−2 with a 5 kcal mol−1 ˚ force constant ensuring consistent sampling, resulting in a 12% ex11 Pursuant to the DOE Public Access Plan, this document represents the authors' peer-reviewed, accepted manuscript. The published version of the article is available from the relevant publisher. change acceptance probability. For farnesene and farnesol bilayer crossing simulations, which were carried out in systems without dodecane solvent present, the [-25,25] ˚A range of the RC was covered by 70 equally spaced umbrellas with a 5 kcal mol−1 ˚A−2 force constant, yielding a 15% exchange acceptance probability. The spacing and force constants are consistent with other umbrella sampling studies of bilayer translocation, 18,27,62–64 although they do result in somewhat lower exchange acceptance probabilities than the 20 % optimum determined from temperature replica exchange studies.65 Extracting farnesene or farnesol into either aqueous solution or a dodecane overlay proceeds in much the same manner, with 55 equally spaced replicas along the [0,650] RC into aqueous solution, and 130 equally spaced replicas along the [-850,650] RC describing the extraction into a dodecane overlay. For both extraction RCs, the applied umbrella used a 0.01 kcal mol−1 contact−2 force constant. For farnesene and farnesol, to convert the eventual permeability coefficient into units that are relevant to experiment, we determine the conversion rate between contact number and distance by a functional fit of the relationship between the biased contact RC and the observed displacement, as was done for the fatty acyl products.18 Since the membrane was observed to become more ordered due to compression of the periodic cell when the organic dodecane phase was present, the center of mass translocations with organic solvent were repeated a second time using a fixed area barostat. During this second set of REUS simulations, the cell dimensions of the system in the membrane plane were fixed to be 60.2 ˚A, consistent with the periodic cell dimensions seen in purely aqueous membrane solvation environments (Fig. S2). Analysis Analysis of the simulation trajectories was primarily carried out using a combination of builtin and purpose built VMD39 scripts and features, leveraging the numpy and scipy libraries66 to handle and manipulate trajectory data. Individual dataplots were created with the matplotlib library,67 and include particle tracking during equilibration, lipid order parameters, 12 Pursuant to the DOE Public Access Plan, this document represents the authors' peer-reviewed, accepted manuscript. The published version of the article is available from the relevant publisher. and naturally also the free energies and diffusivities used to compute permeabilities. Estimation of the free energy profile required for Eq. 1 was performed using a modified version of BayesWHAM,68 which has been accelerated by using Gibbs sampling of the known Dirichlet prior69 rather than the slower Metropolis-Hastings originally implemented.68 The local diffusivity is estimated from the variance and time autocorrelation along the RC, as first formulated by Hummer 70, and used for fatty acyl products previously.18 To account for the exchanges introduced by REUS sampling, the time autocorrelation is determined through an exponential fit to the autocorrelation function at timescales less than the exchange frequency. This approach was also used in determining diffusivity for the fatty acyl products,18 and yields diffusivities in broad agreement with experimentally determined small-molecule diffusivities of 10−6 cm2s−1 for low molecular weight hydrocarbons in an aqueous environment.71,72 The estimates for free energy and diffusivity are then numerically integrated to determine the permeability using the formalism from Eq. 1, with the zero level on the free energy surface taken to be the free energy minima nearest to the level of lipid carbonyls. This means that the permeabilities computed are relative to an already inserted compound in one of the two membrane leaflets. Strains, Media and Chemicals Bacterial Growth Bacterial strains Pseudomonas putida KT2440 and Escherichia coli MG1655 were used for toxicity growth assays. P. putida and E. coli were cultivated in Luria-Bertani (LB) medium (Lennox) containing 10 gL−1 tryptone, 5 gL−1 yeast extract, and 5 gL−1 NaCl for precultures. For growth assays, P. putida and E. coli were cultivated in M9 minimal medium containing 6.78 gL−1 Na2HPO4, 3.00 gL−1 K2HPO4, 0.50 gL−1 NaCl, 1.66 gL−1 NH4Cl, 0.24 gL−1 MgSO4, 0.01 gL−1 CaCl2, and 0.002 gL−1 FeSO4, supplemented with 3.60 gL−1 glucose and/or different concentrations of terpenes. Stock solutions of each terpenoid were prepared as 20% v/v in H20. All the chemicals used for the study were obtained from 13 Pursuant to the DOE Public Access Plan, this document represents the authors' peer-reviewed, accepted manuscript. The published version of the article is available from the relevant publisher. Sigma-Aldrich (St. Louis, MO, USA) or Alfa Aesar (Haverhill, MA, USA- bisabolene only). Yeast Growth The yeast strain Saccharomyces cerevisiae Y294 (ATCC 201160) was used for toxicity growth assays. Pre-cultures of S. cerevisiae were grown in yeast-peptone-dextrose (YPD) liquid medium containing 10 gL−1 yeast extract powder, 20 gL−1 peptone powder, 20 gL−1 glucose. For growth assays, S. cerevisiae was cultivated in yeast nitrogen base (YNB) medium containing 2% (v/v) glucose at pH 4.5 (Sigma Aldrich, St. Louis. MO, USA) containing proper auxotrophic supplements (leucine, uracil, histidine, and tryptophan) at 1 gL−1 and/or different concentrations of terpenes. Growth Assay Toxicity Assays Toxicity of the terpene compounds was evaluated in microplate growth assays performed in a Bioscreen C MBR microplate reader (Growth Curves US, Piscataway, NJ). Pre-cultures were inoculated by scraping strains preserved in 25% v/v glycerol into 5 mL LB (bacteria) or YPD (yeast) medium in 15 mL disposable culture tubes. Pre-culture tubes were incubated shaking at 225 rpm, 30 ◦C. When strains reached OD600 1–2, cells were harvested by centrifugation at 4,500 rpm, and the cell pellets were washed and resuspended in M9 salts without a carbon source. OD600 were measured using a Beckman DU640 spectrophotometer (Beckman Coulter, Brea CA). These resuspended cells were used to inoculate microplate wells containing 150 µL of M9 (bacteria) medium or YNB (yeast) to an OD600 of 0.1. Medium volume of each well was adjusted based on the volume of terpenoid compound required to reach the tested concentration. After strain inoculation, each of the three microbes was grown three times under each of the different conditions tested. These conditions include a control without terpenoid, as well as 0.5%, 1%, and 3% concentrations of bisabolene, bisabolol, farnesene, farnesol, 14 Pursuant to the DOE Public Access Plan, this document represents the authors' peer-reviewed, accepted manuscript. The published version of the article is available from the relevant publisher. limonene, terpineol, and perillyl alcohol. Pinene and its derivatives were not tested due to their comparatively high price. To ensure homogeneity in the inoculum, each terpenoid was vortexed vigorously immediately prior to addition to the microplate well. Microplates were then incubated at 30 ◦C with maximum shaking and growth was measured by reading the absorbance (OD420−580) every 15 min. Growth on Perillyl Alcohol and Terpineol as a Sole Carbon Source We examined the ability for P. putida KT2440 to grow on perillyl alcohol and terpineol as sole carbon sources. Pre-cultures were prepared as above and harvested, washed cells were used to inoculate microplate wells containing 150 µL of M9 medium prepared without glucose to an OD600 of 0.1. Concentrated terpenoids were added to each microplate well as described above. Microplates were then incubated at 30 ◦C with maximum shaking and growth was measured by reading the absorbance (OD420−580) every 15 min. Results System Evolution at Equilibrium In the prior study of fatty acyl products, individual compounds were initially placed within the membrane.18 However, as indicated by Fig. 2, here the compounds were initially placed in solution or at the aqueous/organic interface. The behavior of the individual compounds depends on their chemistry and environment. When initially placed in aqueous solution, as in Fig. 2A, the compounds tend to insert into the bilayer (Fig. 3). The insertion process is not uniformly fast. In some equilibrium trajectories, insertion is not realized prior to the end of 100 ns of equilibration (Fig. S3), whereas others insert rapidly, within the 50 ns timescale previously observed for lipid insertion into a bilayer.57 The observed trend is that the species with the greatest lipophillicity, such as farnesene, limonene, and pinene, are the fastest to insert, with all five 100 ns trajectories leading to a 15 Pursuant to the DOE Public Access Plan, this document represents the authors' peer-reviewed, accepted manuscript. The published version of the article is available from the relevant publisher. 0 10 20 30 40 50 |Compound-Membrane Center-Center Distance| (˚A) 0 Probability Bisabolene Bisabolol Limonene Terpineol Perillyl alcohol Pinene Verbenol Verbenone Farnesene Farnesol Figure 3: Distribution of the position the center of mass of each compound relative to the membrane center over all equilibration trajectories. For each compound, there are two distributions, a solid line for the aqueous equilibration trajectories (Fig. 2A), and a dashed line for the organic equilibration trajectories (Fig. 2B). Similar compounds are grouped by color as in Fig. 1, with more hydrophilic compounds given lighter colors. membrane-embedded compound. The hydroxylated counterparts insert more slowly, with always at least one trajectory failing to insert within 100 ns and remaining in aqueous solution. The compounds also differ in their preferred insertion depth. Hydrogen bonding with lipid carbonyl groups causes the hydroxylated compounds to have a probability peak in the range of 10-15˚A, with the precise value depending on the size of the molecule, consistent with prior simulation at both atomic18,73 and coarse-grained resolution.74 The unmodified terpenoids are all found to be enriched at the membrane center after insertion from aqueous solution (Fig. S3), similar to previous findings for benzene75 and alkanes.18 During equilibration trajectories where a dodecane phase was present, similar to Fig. 2B, spontaneous insertion into the bilayer was rarely observed (Fig. S3). Where insertion into the bilayer from the organic phase does occur, as in bisabolene, limonene, and farnesene, insertion appears to be reversible (Fig. S3). The reversibility of the insertion and the overall population balance between organic and membrane phases suggests that the organic phase may be the preferred partition for these terpenoid compounds. Within the organic phase, compounds are observed to partition based on hydrophilicity. Compounds bearing a hydroxyl 16 Pursuant to the DOE Public Access Plan, this document represents the authors' peer-reviewed, accepted manuscript. The published version of the article is available from the relevant publisher. group prefer to remain on the interface between water and the organic phase, and compounds without potential hydrogen bond interaction sites are nearly uniformly distributed within the dodecane phase (Fig. 3). From these aggregate results, we can already begin to make predictions regarding the overall shape of membrane crossing and extraction free energy surfaces. Oxidized compounds will cross the bilayer more slowly than do the reduced terpenoids, based on the hydrogen bonds created with the phospholipid head groups. Similarly, the oxidized compounds will extract only as far as the dodecane-water interface, and further extraction would bear an energetic penalty. Based on the free diffusion of reduced terpenoids within the organic phase, no additional free energy cost to bury these compounds entirely within the dodecane phase would be expected. Free Energy and Local Diffusivity To test these expectations, as well as to quantify the local free energy and diffusivity such that a permeability can be computed using Eq. 1, we analyze the REUS calculations described previously. However, since farnesene and farnesol use a different RC than the other compounds tested due to the additional rotateable bonds present in their molecular structure (Fig. 1), we must present these results separately. Ring Terpenoids We begin with the majority of the compounds, where the center of mass of the compound relative to the membrane center was used as the RC due to their rings restricting the available degrees of freedom (Fig. 4). The composition of Fig. 4 depends on both aqueous and organic extraction REUS calculations. The free energy and local diffusivity profiles are fused together at the midpoint of the membrane under the assumption that the environment experienced by these small molecules is unaffected by the solvent environment at the membrane periphery. This assumption is not violated so long as the organic extraction is performed in a fixed-area 17 Pursuant to the DOE Public Access Plan, this document represents the authors' peer-reviewed, accepted manuscript. The published version of the article is available from the relevant publisher. −12 −10 −8 −6 −4 −2 0 2 ∆G (kcal/mol) Bisabolene Bisabolol Limonene Terpineol Perillyl alcohol Pinene Verbenol Verbenone −40 −30 −20 −10 0 10 20 30 40 Position Relative to Membrane Center (˚A) 0 5 10 15 20 25 30 35 40 45 D (˚A2/ns) Figure 4: Free energy profile and local diffusivity for compact terpenoid compounds. For the free energy profile, the reference point was chosen to be the compound in aqueous solution (left), which transits the bilayer (center), and is eventually extracted into the organic phase (right). Thus, the RC shown here is a composite of the two REUS calculations, joined at the membrane center (z=0), specifically of the extraction to aqueous solution (left), and the extraction to the organic phase in a fixed area ensemble (right). As a visual aide, a simplified atomistic representation of the respective components is used as a background, with water oxygens shown in blue, dodecane heavy atoms in green, and lipid heavy atoms following a standard color scheme (grey for carbons, red for oxygens, blue for nitrogen, and brown for phosphorus). 18 Pursuant to the DOE Public Access Plan, this document represents the authors' peer-reviewed, accepted manuscript. The published version of the article is available from the relevant publisher. ensemble, as it was for the results presented in Fig. 4. The most straightforward way to test if the presence of the organic solvent interferes with bilayer structure in the membrane interior is to check the profiles for symmetry around the membrane center. When compared with Fig. S4, where the free energy profile is determined from REUS calculations carried out under a semi-isotropic ensemble typically used with membrane simulations, Fig. 4 has significantly more symmetry in the region near the membrane center. The asymmetry with the semi-isotropic barostat comes from lipid ordering induced by the presence of the organic solvent, the origins of which are detailed in a subsequent section. A B Figure 5: Simulation snapshots capturing interfacial behavior of small compounds (green) with both the membrane, water, and dodecane (brown). (A) Bisabolene is relative large, and part of the molecule can interact directly with dodecane while the center of mass is around 20 A. ˚ (B) Pinene is relatively small, and cannot directly interact with dodecane. Instead, in this case the dodecane acts via perturbing water and membrane structure to lower the desorption barrier. The observed barriers for compound extraction from the membrane into either aqueous solution (left) or dodecane (right) are unequal due to differential interaction between the compounds and either water or dodecane. Since the terpenoid compounds are not point 19 Pursuant to the DOE Public Access Plan, this document represents the authors' peer-reviewed, accepted manuscript. The published version of the article is available from the relevant publisher. objects, but rather are extended molecules, larger compounds such as bisabolene may interact directly with dodecane even as their center of mass is within the membrane or near the interface (Fig. 5A). The favorable interactions formed between the dodecane and the terpenoid would naturally yield a lower energetic barrier to desorption than the equivalent extraction into water, which lacks these hydrophobic interactions. However, these smaller barriers also persist in smaller molecules that are too small to directly interact with dodecane while within the membrane headgroup region, such as pinene (Fig. 5B). The reduced organic extraction barrier in small molecules can be ascribed to dodecane ordering the water structure. Aqueous solution is largely bulk, disordered water, where the individual molecules are free to rotate and diffuse, and can find an interaction partner in every direction in space. Near the interface, water molecules preferentially orient to satisfy their hydrogen bonding requirements with nearby lipids, and the impact of this can still be easily observed to 30 ˚A from the bilayer center.76 From our own equilibrium trajectories, we determine the geometric orientation of water molecules as a function of distance from the bilayer center (Fig. S5). Unlike the aqueous simulations, where the ordering breaks down further from the membrane as in previous studies,76 simulations with an organic phase present demonstrate ordering near the water-organic phase interface in addition to the ordering near the membrane. The ordering of the waters near the organic phase reduces the number of easily accessible states, and thereby their entropy with respect to bulk water. The reduced entropy directly lowers the energetic barriers to moving a hydrophobic molecules into solution, since the typical entropy loss associated with creating an ordered water “cage” around the hydrophobic molecule has already happened. The other notable trend from Fig. 4 is the relative extraction barriers between hydroxylated and non-hydroxylated forms of the terpenoids. Hydroxylation uniformly lowers the membrane extraction barrier, with the typical barrier to membrane desorption being reduced by approximately 1 kcal mol−1 relative to the non-hydroxylated product. This would reflect favorable hydrogen bonding between the added hydroxyl and the surrounding wa20 Pursuant to the DOE Public Access Plan, this document represents the authors' peer-reviewed, accepted manuscript. The published version of the article is available from the relevant publisher. ter molecules as the terpenoids enter solution. The trade-off resulting from these hydrogen bonds is that the hydroxylated terpenoids face a 1-4kcal mol−1 larger barrier to crossing the membrane. One final point on the free energies themselves is that the transfer of terpenoid product from the lipid bilayer to the organic phase is always favorable (Fig. 4). This is dissimilar from the fatty acyl results, where transfer from the bilayer to the dodecane phase was typically an energetically unfavorable process, albeit barely.18 The origin of the discrepancy lies in the orientation imposed by the lipid tails. Fatty acyl products are highly similar to lipid acyl tails in their conformational flexibility, and therefore preferentially segregate to the membrane environment rather than the disordered organic phase. By contrast, the rigid terpenoids are rotationally confined within the membrane, but are free to explore these rotational degrees of freedom in the dodecane phase, lowering the free energy there. Similar mechanisms have been previously proposed when aromatic amino acid analogs are transferred into organic solvents within a bilayer.63 The local diffusivities reported in Fig. 4 are all similar in magnitude and shape, reflecting the chemical similarity of the individual compounds. Based on the local variation in the diffusivity (Fig. 4), we estimate the uncertainty of individual values to be below 10 %, which will have only a small effect on the computed permeability. The shape of this shared profile reflects the local structure of the medium being traversed. In regions of disorder, such as in solution or the organic phase, the local diffusion is highest, with aqueous solution demonstrating higher diffusion than the organic phase due to the faster diffusion of water relative to dodecane. Similarly, near the membrane center where the tails are the least ordered, there is another increase in diffusivity (Fig. 4) similar to that seen in some small molecules,77–79 as well as the comparison fatty acid simulations.18 The mechanism for these regions of high diffusion may be voids at the interface between leaflets formed by tail movement. The reduced pressure environment of the microscopic voids would draw in terpenoids, raising the diffusion relative to the remainder of the membrane, where molecular collisions with the 21 Pursuant to the DOE Public Access Plan, this document represents the authors' peer-reviewed, accepted manuscript. The published version of the article is available from the relevant publisher. dense lipid tails retards motion.79 Farnesene and Farnesol Figure 6: Free energy profile and local diffusivity for farnesene and farnesol membrane crossing, as well as their nearest equivalent fatty acyl products previously studied18 for comparison. Farnesene and farnesol were treated separately from the other terpenoids considered due to the different RCs required to describe membrane crossing and extraction. For membrane crossing, only the position of the C1 carbon was used to drive the transition from one membrane leaflet to the other (Fig. 6), analogous to the RC used for prior fatty acid studies,18 making these previous results the natural point of comparison. The free energy profiles and diffusivities for farnesene and farnesol show the same chemistry-dependent trends as the 22 Pursuant to the DOE Public Access Plan, this document represents the authors' peer-reviewed, accepted manuscript. The published version of the article is available from the relevant publisher. equivalent fatty acyl products. Farnesene prefers to be at the membrane center, similar to the other hydrophobic terpenoids, whereas farnesol exhibits a small crossing barrier similar to the other terpene alcohols (Fig. 4). The measured diffusivity for farnesene and farnesol crossings are also in line with the other terpenoids, with a diffusion rate near the bilayer center A2 of around 20 ˚ ns−1, which is less than the previously reported diffusivities for comparable fatty acid products (Fig. 6). Mechanistically, the extra unsaturation sites in the farnesene rigidify the molecule, preventing farnesene from easily filling in voids within the membrane that do not have the right shape, slowing diffusion relative to the comparatively flexible saturated fatty acids. Contact Number −5 0 5 10 15 ∆G (kcal/mol) A Pentadecane Cetyl alcohol Farnesene Farnesol Contact Number B 0 100 200 300 400 500 600 Contact Number 0 2 4 6 8 D (Contact Number 2/fs) −800 −600 −400 −200 0 200 400 600 Contact Number Figure 7: Free energy profile and local diffusivity for farnesene and farnesol membrane extraction into aqueous solution (A) and into an organic phase (B). For comparison, the nearest equivalent fatty acyl products previously studied,18 cetyl alcohol and pentadecane, are also included. The definition for the contact number is provided in Eq. 2, keeping in mind that extraction from the membrane proceeds from right to left as the contact number is decreased. Extraction of farnesene and farnesol into both aqueous (Fig. 7A) and organic (Fig. 7B) environments also exhibit existing trends observed for the other terpenoid compounds tested. The barrier to extraction into both environments is lower for farnesol, just as it is for fatty 23 Pursuant to the DOE Public Access Plan, this document represents the authors' peer-reviewed, accepted manuscript. The published version of the article is available from the relevant publisher. alcohols or hydroxylated terpenoids more generally. Following the same trend as the other terpenoids, farnesene and farnesol have a lower free energy in dodecane where rotational degrees of freedom for the compound as a whole are not constrained as they would be within the bilayer. The diffusivities are broadly in line with what was observed for the fatty acyl products, although no direct comparison can be made due to the different units that the diffusion was measured in relative to the crossing step or that of the other terpenoids more broadly. To handle the different units for diffusivity in Eq. 1, we determine the relationship between contact number and displacement in Fig. S6. Induced Lipid Ordering by Water/Organic Interfaces Before determining the permeabilities, we revisit the impact of the water-dodecane interface on membrane dynamics, which necessitated moving to a constant area simulation barostat to mitigate asymmetries in the internal bilayer free energy profile (Fig. S4). In the original design of the simulations, the significant surface tension generated at the water-dodecane interface80 and its coupling to membrane structure was not appreciated. Typical surface tensions for membrane components and water are between 1-5 mN m−1, 81 while those for the dodecane water interface are around 50 mN m−1. 80 This means that the preferred contact area between water and dodecane will be smaller than the contact area between water and membrane. In a periodic simulation system, the shrinking water-dodecane interface will therefore compress the membrane-water interface. The coupling of the two interfaces in a conventional barostat will therefore shrink both the membrane and doecane-water interfacial surface areas (Fig. S2), resulting in a thicker membrane. We highlight here that the membrane thickening is unnatural. In an experimental setting, the two interfacial tensions would not be coupled as they are in a periodic membrane simulation, and so there is no reason to suspect membranes thicken in the presence of an organic solvent. The thickening of the bilayer changed the behavior of the individual membrane lipids as well. Specifically, the thicker bilayer packed together individual lipid tails more closely, 24 Pursuant to the DOE Public Access Plan, this document represents the authors' peer-reviewed, accepted manuscript. The published version of the article is available from the relevant publisher. Figure 8: Deuterium order parameters for each of the prominent lipid tails compared between aqueous (black) and organic (gray) systems during unbiased equilibrium simulation in a typical constant ratio barostat. 25 Pursuant to the DOE Public Access Plan, this document represents the authors' peer-reviewed, accepted manuscript. The published version of the article is available from the relevant publisher. resulting in a more ordered bilayer (Fig. 8). Small molecules have been previously shown to partition between coexisting ordered and disordered membrane phases depending on their specific chemistry.82–84 Thus, the free energy profile changes between Figs. 4 and S4 are purely an artifact of lipid ordering induced by the dodecane-water interfacial surface tension if the barostat is allowed to control all dimensions of the periodic cell. Since the permeabilities are exponentially dependent on the free energies (Eq. 1), these free energy differences will significantly impact the result. In the previous fatty acyl product study,18 as well as the aqueous crossings performed in this work, no additional surface tension was present that would thicken the bilayer, since these calculations were performed in the absence of an organic layer. To mimic this, we use a fixed area barostat for the organic simulations involved in generating Fig. 4 to fix the membrane width, and thereby preserve membrane internal structure. Permeability Having determined all the requirements to compute a permeability based on Eq. 1, we can now tabulate the permeabilities of each step (Table 1), which in turn are proportional to the anticipated flux for a given concentration differential.18 The overall permeability for the total crossing and extraction process (Pt) is limited by the slower of the crossing (Pc) and extraction (Pe) permeabilities due to the additive nature of Eq. 1: ⎧ Pc if Pc << Pe ⎪⎪⎪⎪⎪⎪⎨ 1 PcPe Pt = = ' Pe if Pe << Pc (3) P −1 + P −1 c e Pc + Pe ⎪⎪⎪⎪⎪⎪⎩ 1 2Pc if Pe = Pc The implication from Eq. 3 is that ideally Pc and Pe would be approximately equal to maximize the overall permeability. For terpenoids, we observe that the permeability for crossing is faster than that of the 26 Pursuant to the DOE Public Access Plan, this document represents the authors' peer-reviewed, accepted manuscript. The published version of the article is available from the relevant publisher. Table 1: Permeability table. Individual columns estimate the permeability of crossing the bilayer from one inserted energy minimum to another (P), the extraction permeability of going from the bilayer into aqueous solution (Paq), and the extraction permeability of going from the bilayer into the dodecane phase (Pphase). Pentadecane and cetyl alcohol permeabilities were reanalyzed using BayesWHAM,68 which resulted in small changes to some of the previously reported permeabilities.18 Product Bilayer Crossing log10 P � cm s Extraction log10 Paq � cm s log10 Pphase � cm s Bisabolene Bisabolol 1.2 0.6 -5.4 -4.0 0.6 1.2 Limonene Terpineol Perillyl alcohol 1.0 -0.0 -1.1 -3.5 -1.8 -2.2 -0.2 0.2 -1.4 Pinene Verbenol Verbenone 1.2 -0.2 0.2 -3.5 -1.5 -1.2 0.5 0.0 1.5 Farnesene Farnesol 1.6 -0.1 -4.3 -2.2 -0.2 0.5 Pentadecane Cetyl alcohol 1.2 -1.3 -5.3 -3.2 -2.3 -1.6 27 Pursuant to the DOE Public Access Plan, this document represents the authors' peer-reviewed, accepted manuscript. The published version of the article is available from the relevant publisher. extraction into the organic phase, similar to what had been noted for alkanes or alkenes.18 In isolation, this suggests that there may be a benefit to making the compounds more hydrophilic so that they extract more easily. However, when a hydroxyl group is added to the terpenoids, the crossing is slowed sufficiently so that it then becomes the rate-limiting step of the overall extraction process. The bisabolene/bisabolol pairing demonstrates this the most clearly, where the bilayer crossing permeabilities and organic phase extraction permeabilities are swapped. By contrast, earlier results for alkanes indicated that the extraction and crossing permeabilities were balanced for the corresponding alcohol (Table 1), resulting in the fastest possible Pt. Thus, for the terpenoids considered in their best case scenario for extraction (adjacent organic and membrane phases), the total extraction permeability is effectively unchanged upon oxidation. If however, the terpenoids are forced to be extracted via aqueous solution, there is a significant gain in the total permeability. Uniformly, Paq << Pc in Table 1, with Paq then becoming the rate limiting step for the total permeability (Eq. 3). Since extraction directly into aqueous solution is 10-100 times faster for hydroxylated terpenoids due to the lower free energy barriers to extraction into solution (Figs. 4 and 7A), oxidation in this case would increase the rate of product release from the host cells. Terpenoid Toxicity Assessment In addition to permeability, compound toxicity is an important metric through which we can measure the suitability of the target terpenoid compounds for industrial applications. Due to the high partition coefficients for the compound studied (Table S1, based on Figs. 4 and 7), the terpenoids will partition strongly into the membrane, with enrichments of up to 1010 relative to the concentrations in solution. Thus, the compounds will be found in high concentrations within the membrane under typical conditions. As was already demonstrated in Fig. 3, the oxidation of the terpenoids controls the location of partitioning, with oxidized terpenoids found more frequently at the membrane surface rather than within the membrane 28 Pursuant to the DOE Public Access Plan, this document represents the authors' peer-reviewed, accepted manuscript. The published version of the article is available from the relevant publisher. core. The differential accessibility of the terpenoids suggests that there may be two different toxicity mechanisms, depending on the species. Hydrophobic compounds could accumulate within the bilayer center, causing the membrane to swell and potentially rupture at very high concentrations. Conversely, hydrophilic compounds would cause membrane thinning as they increase the membrane surface area, or alternatively may act as an agonist to membraneassociated proteins and inhibit their function. Testing different species under physiologically relevant concentrations of these compounds permits us to discriminate which mechanism is more problematic for industrial applications. Assaying toxicity effects of these various compounds in organisms with varying robustness can also inform host-target pair selection for an industrial process. The two most commonly used organisms for terpene production were included, Saccharomyces cerevisiae and E. coli, as well as the addition of the stress-tolerant and organic solvent-tolerant Pseudomonas putida KT2440. In one report, the toxicity of the terpenoid geranic acid was assayed against E. coli, S. cerevisiae, and P. putida. 85 Growth of E. coli and S. cerevisiae was inhibited at 7 mM and 2 mM, respectively, whereas the growth of P. putida was not inhibited at the highest concentration tested, 40 mM.85 We therefore sought to identify any toxicity effects between the oxidized and non-oxidized terpenoids in E. coli, S. cerevisiae, and P. putida. For aqueous media containing up to 3% terpenoid product, representing a high titer environment for the cell, generally both oxidized and non-oxidized terpenoids were tolerated (Fig. 9). However, the tolerance was not uniform between products, with oxidized limonene derivatives in particular strongly retarding growth for yeast and E. coli. Other oxidized terpenoids caused growth to lag behind the rapid growth observed in non-oxidized terpenoids, which indicates that the strain maintains some tolerance to these compounds. Interestingly, in some instances cell density increases in the presence of oxidized products, suggesting that they may have been incorporated into cellular carbon metabolism, demonstrated conclusively by growth when the terpenoids are the sole carbon source (Fig. S7). 29 Pursuant to the DOE Public Access Plan, this document represents the authors' peer-reviewed, accepted manuscript. The published version of the article is available from the relevant publisher. 0.0 0.5 1.0 Bisabolene Bisabolol No additive Limonene Terpineol Perillyl alcohol No additive S. cerevisiae Farnesene Farnesol No additive 0.0 0.5 1.0 P. putida 0 10 20 30 40 0.0 0.5 1.0 10 20 30 40 10 20 30 40 E. coli Optical Density Time (hours) Figure 9: Microbial growth dependence on environmental terpenoid concentration. Each of the three microbial species (S. cerevisiae, P. putida, and E. coli) were exposed to different concentrations of readily available terpenoid products, indicated in the in-figure legends. The 0.5% concentration results are drawn as solid lines, the 1% concentrations are dashed lines, and 3% terpenoid concentration results are dotted lines. The drawn lines represent the mean of three replicates, with typical deviations of 5-10% around the mean value, which are not shown to improve the clarity of the figure. 30 Pursuant to the DOE Public Access Plan, this document represents the authors' peer-reviewed, accepted manuscript. The published version of the article is available from the relevant publisher. These results indicate that under some conditions a hydrophilic membrane thinning or interference mechanism is likely at work. However, since both oxidized and non-oxidized terpenoid variants do show growth even under significant concentrations of product, the results suggest that product toxicity is likely only a marginal concern for industrial terpenoid production, and that neither of the potential toxicity mechanisms are a substantial concern in the biological production of terpenes. Discussion With the computational results coupled to experiment indicating that growth proceeds for all terpenoids tested, we can now address the hypothesis tested in this work, namely that oxidation may improve product extraction rates. When combined, the available evidence suggests that the impact of hydroxylation on terpenoid extraction rates depends on the experimental conditions. If there is little dodecane relative to the aggregate cell volume, and those cells are widely separated such that there is ample solution between them as in a low density laboratory culture, the aqueous extraction permeabilities from Table 1 would better represent system kinetics, with the aqueous extraction step being rate limiting. In this case, oxidation of the terpenoids within the cell would be expected to dramatically increase the rate at which terpenoid product can accumulate into the dodecane phase. However, in an abundance of dodecane, or perhaps where cells are close enough in a dense industrial culture that terpenoid products can easily transit from cell to cell prior to eventual extraction into dodecane, the extraction into organic phase results will be more applicable. An informative exercise is to compute the relative surface areas of different components such that equal flux will exit the cell via both the largely aqueous and the organic-adjacent pathways. If we compute Pt using Eq. 3 assuming that Pe is either Paq or Pphase, we find that the equal flux condition for non-hydroxylated terpenes occurs when only around one hundredth of one percent (0.01 %) of the total cellular surface area acts like the close dodecane31 Pursuant to the DOE Public Access Plan, this document represents the authors' peer-reviewed, accepted manuscript. The published version of the article is available from the relevant publisher. membrane contact situation shown in Fig. 2B rather than the aqueous extraction condition (Fig. 2A). For oxidized terpenoids, the cutoff percentage is significantly higher, with direct membrane-dodecane extraction being predominant only if the contact area makes up more than 1 % of the total cell surface. How much cell area is actually in close proximity to dodecane? If we use a hard-sphere approximation, and assume again that dodecane is large enough to constitute a plane adjacent to the cell, we determine that only .5 % of a bacteria-sized sphere (380 nm radius) would be in close enough contact to extract directly from the membrane to the dodecane phase. As the cell size increases to that of a typical yeast cell (2500 nm radius), only 0.06 % of the cell surface would make close contact with dodecane, suggesting that the aqueous extraction would constitute the majority of the flux under these conditions. The implication of these calculations is that if the environmental conditions are such that cell extraction always proceeds directly from membrane to dodecane without an aqeuous intermediate, oxidation of terpenoids will not actually improve overall membrane permeability. However, if the probability of close contact between the membrane and dodecane phase is low, as may be the case due to cellular avoidance of dodecane phases26 or the simple low density of laboratory suspended cultures, oxidizing terpenoids would increase terpenoid product flux through aqueous intermediate states significantly, thereby increasing the overall flux into the organic overlay. Conclusion The initial hypothesis was that hydroxylating terpenoids would accelerate the product extraction step relative to the typical lipophilic terpenoid products. Terpenoid oxidation does indeed increase the rate of product extraction, while simultaneously lowering the membrane crossing rate of the compounds. Since the rates of crossing and extraction for conventional terpenoids were already well balanced, however, there is little net effect on the overall per32 Pursuant to the DOE Public Access Plan, this document represents the authors' peer-reviewed, accepted manuscript. The published version of the article is available from the relevant publisher. meation rate of the designed products entering a proximal dodecane phase. Thus, unlike previous results for fatty acyl products,18 the value proposition of choosing one specific degree of oxidation is not clear outside of low density laboratory cultures, as other metrics also may contribute to the optimal solution. One such metric is toxicity, which we demonstrate is largely indifferent to hydroxylation of terpenoids. Oxidation did delay or impair microbial growth in some instances, but typically was not toxic, just like the non-oxidized terpenoids. Instead, extended analysis of the fluxes to the organic phase suggest that the most important metric in determining if oxidation of terpenoids would improve extraction productivity depends on the aqueous volume between the cell and the organic phase. The measured extraction rates to the organic phase are accelerated by ordered water at the narrow gap between membrane and dodecane, such that high-density cultures where cells make contact at the atomic scale are indifferent to terpenoid oxidation. However, if the dodecane overlay only rarely contacts the cell membrane, as in many low-density laboratory cultures, improving the terpenoid solubility by adding a hydroxyl would significantly improve the overall terpenoid production rate. Determining the precise cell density and geometry demarcating the border between these two regimes is beyond the scope of current research, and may be a direction for future study. Acknowledgement We thank Alan Grossfield for fruitful discussion on the molecular origins of membrane perturbation by the addition of an organic phase. We also acknowledge the insight provided by Lahiru Jayakody into possible metrics for assessing toxicity within microbial systems. This work was authored by Alliance for Sustainable Energy, LLC, the manager and operator of the National Renewable Energy Laboratory for the U.S. Department of Energy (DOE) under Contract No. DE-AC36-08GO28308. JVV is supported by the NREL Director’s Fellowship funded by the Laboratory Directed Research and Development (LDRD) program. GJB, 33 Pursuant to the DOE Public Access Plan, this document represents the authors' peer-reviewed, accepted manuscript. The published version of the article is available from the relevant publisher. GTB, and MFC are supported by the Department of Energy Office of Energy Efficiency and Renewable Energy Bioenergy Technologies Office. A portion of the research was performed using computational resources sponsored by the Department of Energy’s Office of Energy Efficiency and Renewable Energy and located at the National Renewable Energy Laboratory. This work also used the Extreme Science and Engineering Discovery Environment (XSEDE)86 through grant TG-MCB090159 to G.T.B., specifically the Stampede2 system at the University of Texas at Austin in the Texas Advanced Computing Center. XSEDE is supported by National Science Foundation grant number ACI-1548562. Supporting Information Available The Supporting Information pdf contains ancillary tables and figures related to discussion within the main text. References (1) George, K. W.; Alonso-Gutierrez, J.; Keasling, J. D.; Lee, T. S. In Biotechnology of Isoprenoids; Schrader, J., Bohlmann, J., Eds.; Springer International Publishing: Cham, 2015; pp 355–389. (2) Mewalal, R.; Rai, D. K.; Kainer, D.; Chen, F.; K¨ulheim, C.; Peter, G. F.; Tuskan, G. A. Plant-Derived Terpenes: A Feedstock for Specialty Biofuels. Trends in Biotechnology 2017, 35, 227–240. (3) Figueiredo, A. C.; Barroso, J. G.; Pedro, L. G.; Scheffer, J. J. C. Factors Affecting Secondary Metabolite Production in Plants: Volatile Components and Essential Oils. Flavour and Fragrance Journal 2008, 23, 213–226. (4) Martin, V. J. J.; Pitera, D. J.; Withers, S. T.; Newman, J. D.; Keasling, J. D. Engi34 Pursuant to the DOE Public Access Plan, this document represents the authors' peer-reviewed, accepted manuscript. The published version of the article is available from the relevant publisher. neering a Mevalonate Pathway in Escherichia Coli for Production of Terpenoids. Nature Biotechnology 2003, 21, 796–802. (5) Schempp, F. M.; Drummond, L.; Buchhaupt, M.; Schrader, J. Microbial Cell Factories for the Production of Terpenoid Flavor and Fragrance Compounds. Journal of Agricultural and Food Chemistry 2018, 66, 2247–2258. (6) Davies, F. K.; Jinkerson, R. E.; Posewitz, M. C. Toward a Photosynthetic Microbial Platform for Terpenoid Engineering. Photosynthesis Research 2015, 123, 265–284. (7) Kampranis, S. C.; Makris, A. M. Developing a Yeast Cell Factory for the Production of Terpenoids. Computational and Structural Biotechnology Journal 2012, 3, e201210006. (8) Zhang, H.; Liu, Q.; Cao, Y.; Feng, X.; Zheng, Y.; Zou, H.; Liu, H.; Yang, J.; Xian, M. Microbial Production of Sabinene—A New Terpene-Based Precursor of Advanced Biofuel. Microbial Cell Factories 2014, 13, 20. (9) Leonard, E.; Ajikumar, P. K.; Thayer, K.; Xiao, W.-H.; Mo, J. D.; Tidor, B.; Stephanopoulos, G.; Prather, K. L. J. Combining Metabolic and Protein Engineering of a Terpenoid Biosynthetic Pathway for Overproduction and Selectivity Control. Proceedings of the National Academy of Sciences 2010, 107, 13654–13659. (10) Peralta-Yahya, P. P.; Ouellet, M.; Chan, R.; Mukhopadhyay, A.; Keasling, J. D.; Lee, T. S. Identification and Microbial Production of a Terpene-Based Advanced Biofuel. Nature Communications 2011, 2, 483. (11) Tholl, D. Terpene Synthases and the Regulation, Diversity and Biological Roles of Terpene Metabolism. Current Opinion in Plant Biology 2006, 9, 297–304. (12) Gershenzon, J.; Dudareva, N. The Function of Terpene Natural Products in the Natural World. Nature Chemical Biology 2007, 3, 408–414. 35 Pursuant to the DOE Public Access Plan, this document represents the authors' peer-reviewed, accepted manuscript. The published version of the article is available from the relevant publisher. (13) Trombetta, D.; Castelli, F.; Sarpietro, M. G.; Venuti, V.; Cristani, M.; Daniele, C.; Saija, A.; Mazzanti, G.; Bisignano, G. Mechanisms of Antibacterial Action of Three Monoterpenes. Antimicrobial Agents and Chemotherapy 2005, 49, 2474–2478. (14) Dorman, H. J. D.; Deans, S. G. Antimicrobial Agents From Plants: Antibacterial Activity of Plant Volatile Oils. Journal of Applied Microbiology 2000, 88, 308–316. (15) Kang, M.-K.; Eom, J.-H.; Kim, Y.; Um, Y.; Woo, H. M. Biosynthesis of Pinene From Glucose Using Metabolically-Engineered Corynebacterium Glutamicum. Biotechnology Letters 2014, 36, 2069–2077. (16) Schalk, M.; Pastore, L.; Mirata, M. A.; Khim, S.; Schouwey, M.; Deguerry, F.; Pineda, V.; Rocci, L.; Daviet, L. Toward a Biosynthetic Route to Sclareol and Amber Odorants. Journal of the American Chemical Society 2012, 134, 18900–18903. (17) Brennan, T. C.; Turner, C. D.; Kr¨omer, J. O.; Nielsen, L. K. Alleviating Monoterpene Toxicity Using a Two-Phase Extractive Fermentation for the Bioproduction of Jet Fuel Mixtures in Saccharomyces Cerevisiae. Biotechnology and Bioengineering 2012, 109, 2513–2522. (18) Vermaas, J. V.; Beckham, G. T.; Crowley, M. F. Membrane Permeability of Fatty Acyl Compounds Studied via Molecular Simulation. The Journal of Physical Chemistry B 2017, 121, 11311–11324. (19) Zada, B.; Wang, C.; Park, J.-B.; Jeong, S.-H.; Park, J.-E.; Singh, H. B.; Kim, S.-W. Metabolic Engineering of Escherichia Coli for Production of Mixed Isoprenoid Alcohols and Their Derivatives. Biotechnology for Biofuels 2018, 11, 210. (20) Wang, C.; Kim, J.-Y.; Choi, E.-S.; Kim, S.-W. Microbial Production of Farnesol (FOH): Current States and Beyond. Process Biochemistry 2011, 46, 1221–1229. 36 Pursuant to the DOE Public Access Plan, this document represents the authors' peer-reviewed, accepted manuscript. The published version of the article is available from the relevant publisher. (21) Han, G. H.; Kim, S. K.; Yoon, P. K.-S.; Kang, Y.; Kim, B. S.; Fu, Y.; Sung, B. H.; Jung, H. C.; Lee, D.-H.; Kim, S.-W. et al. Fermentative Production and Direct Extraction of (-)-α-bisabolol in Metabolically Engineered Escherichia Coli. Microbial Cell Factories 2016, 15, 185. (22) Rottava, I.; Cortina, P. F.; Zanella, C. A.; Cansian, R. L.; Toniazzo, G.; Treichel, H.; Antunes, O. A. C.; Oestreicher, E. G.; de Oliveira, D. Microbial Oxidation of (-)-αpinene to Verbenol Production by Newly Isolated Strains. Applied Biochemistry and Biotechnology 2010, 162, 2221–2231. (23) Agrawal, R.; Joseph, R. Bioconversion of Alpha Pinene to Verbenone by Resting Cells of Aspergillus Niger. Applied Microbiology and Biotechnology 2000, 53, 335–337. (24) Adams, A.; Demyttenaere, J. C.; De Kimpe, N. Biotransformation of (R)-(+)- And (S)- (-)-Limonene to α-terpineol by Penicillium Digitatum— Investigation of the Culture Conditions. Food Chemistry 2003, 80, 525–534. (25) Alonso-Gutierrez, J.; Chan, R.; Batth, T. S.; Adams, P. D.; Keasling, J. D.; Petzold, C. J.; Lee, T. S. Metabolic Engineering of Escherichia Coli for Limonene and Perillyl Alcohol Production. Metabolic Engineering 2013, 19, 33–41. (26) Kleinegris, D. M. M.; van Es, M. A.; Janssen, M.; Brandenburg, W. A.; Wijffels, R. H. Phase Toxicity of Dodecane on the Microalga Dunaliella Salina. Journal of Applied Phycology 2011, 23, 949–958. (27) Lee, C. T.; Comer, J.; Herndon, C.; Leung, N.; Pavlova, A.; Swift, R. V.; Tung, C.; Rowley, C. N.; Amaro, R. E.; Chipot, C. et al. Simulation-Based Approaches for Determining Membrane Permeability of Small Compounds. Journal of Chemical Information and Modeling 2016, 56, 721–733. (28) Sugita, Y.; Kitao, A.; Okamoto, Y. Multidimensional Replica-Exchange Method for Free-Energy Calculations. Journal of Chemical Physics 2000, 113, 6042–6051. 37 Pursuant to the DOE Public Access Plan, this document represents the authors' peer-reviewed, accepted manuscript. The published version of the article is available from the relevant publisher. (29) Diamond, J. M.; Katz, Y. Interpretation of Nonelectrolyte Partition Coefficients Between Dimyristoyl Lecithin and Water. The Journal of Membrane Biology 1974, 17, 121–154. (30) Marrink, S.-J.; Berendsen, H. J. C. Simulation of Water Transport Through a Lipid Membrane. Journal of Physical Chemistry 1994, 98, 4155–4168. (31) Daum, G.; Tuller, G.; Nemec, T.; Hrastnik, C.; Balliano, G.; Cattel, L.; Milla, P.; Rocco, F.; Conzelmann, a.; Vionnet, C. et al. Systematic Analysis of Yeast Strains With Possible Defects in Lipid Metabolism. Yeast (Chichester, England) 1999, 15, 601–614. (32) Lindberg, L.; Santos, A. X. S.; Riezman, H.; Olsson, L.; Bettiga, M. Lipidomic Profiling of Saccharomyces Cerevisiae and Zygosaccharomyces Bailii Reveals Critical Changes in Lipid Composition in Response to Acetic Acid Stress. PLoS ONE 2013, 8, e73936. (33) Ejsing, C. S.; Sampaio, J. L.; Surendranath, V.; Duchoslav, E.; Ekroos, K.; Klemm, R. W.; Simons, K.; Shevchenko, A. Global Analysis of the Yeast Lipidome by Quantitative Shotgun Mass Spectrometry. Proc. Natl. Acad. Sci. U.S.A. 2009, 106, 2136–2141. (34) Fei, Q.; O’Brien, M.; Nelson, R.; Chen, X.; Lowell, A.; Dowe, N. Enhanced Lipid Production by Rhodosporidium Toruloides Using Different Fed-Batch Feeding Strategies With Lignocellulosic Hydrolysate as the Sole Carbon Source. Biotechnology for Biofuels 2016, 9, 130. (35) Li, Y.; Zhao, Z. K.; Bai, F. High-Density Cultivation of Oleaginous Yeast Rhodosporidium Toruloides Y4 in Fed-Batch Culture. Enzyme and Microbial Technology 2007, 41, 312–317. (36) Jo, S.; Kim, T.; Iyer, V. G.; Im, W. CHARMM-GUI: A Web-Based Graphical User Interface for CHARMM. Journal of Computational Chemistry 2008, 29, 1859–1865. 38 Pursuant to the DOE Public Access Plan, this document represents the authors' peer-reviewed, accepted manuscript. The published version of the article is available from the relevant publisher. (37) Jo, S.; Lim, J. B.; Klauda, J. B.; Im, W. CHARMM-GUI Membrane Builder for Mixed Bilayers and Its Application to Yeast Membranes. Biophysical Journal 2009, 97, 50–58. (38) Wu, E. L.; Cheng, X.; Jo, S.; Rui, H.; Song, K. C.; D´avila-Contreras, E. M.; Qi, Y.; Lee, J.; Monje-Galvan, V.; Venable, R. M. et al. CHARMM-GUI Membrane Builder Toward Realistic Biological Membrane Simulations. Journal of Computational Chemistry 2014, 35, 1997–2004. (39) Humphrey, W.; Dalke, A.; Schulten, K. VMD: Visual Molecular Dynamics. Journal of Molecular Graphics 1996, 14, 33–38. (40) Mashl, R. J.; Scott, H. L.; Subramaniam, S.; Jakobsson, E. Molecular Simulation of Dioleoylphosphatidylcholine Lipid Bilayers at Differing Levels of Hydration. Biophysical Journal 2001, 81, 3005–3015. (41) H¨ogberg, C.-J.; Lyubartsev, A. P. A Molecular Dynamics Investigation of the Influence of Hydration and Temperature on Structural and Dynamical Properties of a Dimyristoylphosphatidylcholine Bilayer. The Journal of Physical Chemistry B 2006, 110, 14326–14336. (42) Klauda, J. B.; Venable, R. M.; Freites, J. A.; O’Connor, J. W.; Tobias, D. J.; Mondragon-Ramirez, C.; Vorobyov, I.; MacKerell, A. D.; Pastor, R. W. Update of the CHARMM All-Atom Additive Force Field for Lipids: Validation on Six Lipid Types. The Journal of Physical Chemistry. B 2010, 114, 7830–7843. (43) Phillips, J. C.; Braun, R.; Wang, W.; Gumbart, J.; Tajkhorshid, E.; Villa, E.; Chipot, C.; Skeel, R. D.; Kal´e, L.; Schulten, K. Scalable Molecular Dynamics With NAMD. Journal of Computational Chemistry 2005, 26, 1781–1802. (44) Lim, J. B.; Rogaski, B.; Klauda, J. B. Update of the Cholesterol Force Field Parameters in CHARMM. Journal of Physical Chemistry B 2012, 116, 203–210. 39 Pursuant to the DOE Public Access Plan, this document represents the authors' peer-reviewed, accepted manuscript. The published version of the article is available from the relevant publisher. (45) Vanommeslaeghe, K.; Hatcher, E.; Acharya, C.; Kundu, S.; Zhong, S.; Shim, J.; Darian, E.; Guvench, O.; Lopes, P.; Vorobyov, I. et al. CHARMM General Force Field: A Force Field for Drug-Like Molecules Compatible With the CHARMM All-Atom Additive Biological Force Fields. Journal of Computational Chemistry 2010, 31, 671–690. (46) Vanommeslaeghe, K.; MacKerell, A. D. Automation of the CHARMM General Force Field (CGenFF) I: Bond Perception and Atom Typing. Journal of Chemical Information and Modeling 2012, 52, 3144–3154. (47) Jorgensen, W. L.; Chandrasekhar, J.; Madura, J. D.; Impey, R. W.; Klein, M. L. Comparison of Simple Potential Functions for Simulating Liquid Water. The Journal of Chemical Physics 1983, 79, 926. (48) Darden, T.; York, D.; Pedersen, L. Particle Mesh Ewald: An N Log(N) Method for Ewald Sums in Large Systems. The Journal of Chemical Physics 1993, 98, 10089– 10092. (49) Essmann, U.; Perera, L.; Berkowitz, M. L.; Darden, T.; Lee, H.; Pedersen, L. G. A Smooth Particle Mesh Ewald Method. The Journal of Chemical Physics 1995, 103, 8577. (50) Kubo, R. The Fluctuation-Dissipation Theorem. Reports on Progress in Physics 1966, 29, 255–284. (51) Feller, S. E.; Zhang, Y.; Pastor, R. W.; Brooks, B. R. Constant Pressure Molecular Dynamics Simulation: The Langevin Piston Method. The Journal of Chemical Physics 1995, 103, 4613–4621. (52) Martyna, G. J.; Tobias, D. J.; Klein, M. L. Constant Pressure Molecular Dynamics Algorithms. The Journal of Chemical Physics 1994, 101, 4177. 40 Pursuant to the DOE Public Access Plan, this document represents the authors' peer-reviewed, accepted manuscript. The published version of the article is available from the relevant publisher. (53) Miyamoto, S.; Kollman, P. a. SETTLE: An Analytical Version of the SHAKE and RATTLE Algorithm for Rigid Water Models. Journal of Computational Chemistry 1992, 13, 952–962. (54) Hong, C.; Tieleman, D. P.; Wang, Y. Microsecond Molecular Dynamics Simulations of Lipid Mixing. Langmuir 2014, 30, 11993–12001. (55) Bera, I.; Klauda, J. B. Molecular Simulations of Mixed Lipid Bilayers With Sphingomyelin, Glycerophospholipids, and Cholesterol. The Journal of Physical Chemistry B 2017, 121, 5197–5208. (56) Ing´olfsson, H. I.; Melo, M. N.; van Eerden, F. J.; Arnarez, C.; L´opez, C. A.; Wassenaar, T. A.; Periole, X.; De Vries, A. H.; Tieleman, D. P.; Marrink, S. J. Lipid Organization of the Plasma Membrane. Journal of the American Chemical Society 2014, 136, 14554–14559. (57) Vermaas, J. V.; Tajkhorshid, E. A Microscopic View of Phospholipid Insertion Into Biological Membranes. Journal of Physical Chemistry B 2014, 118, 1754–1764. (58) Tahir, M. A.; Van Lehn, R. C.; Choi, S. H.; Alexander-Katz, A. Solvent-Exposed Lipid Tail Protrusions Depend on Lipid Membrane Composition and Curvature. Biochimica Et Biophysica Acta - Biomembranes 2016, 1858, 1207–1215. (59) Kuˇcerka, N.; Nagle, J. F.; Sachs, J. N.; Feller, S. E.; Pencer, J.; Jackson, A.; Katsaras, J. Lipid Bilayer Structure Determined by the Simultaneous Analysis of Neutron and XRay Scattering Data. Biophysical Journal 2008, 95, 2356–2367. (60) Fiorin, G.; Klein, M. L.; H´enin, J. Using Collective Variables to Drive Molecular Dynamics Simulations. Molecular Physics 2013, 111, 3345–3362. (61) Sindhikara, D. J.; Emerson, D. J.; Roitberg, A. E. Exchange Often and Properly in 41 Pursuant to the DOE Public Access Plan, this document represents the authors' peer-reviewed, accepted manuscript. The published version of the article is available from the relevant publisher. Replica Exchange Molecular Dynamics. Journal of Chemical Theory and Computation 2010, 6, 2804–2808. (62) MacCallum, J. L.; Bennett, W. F. D.; Tieleman, D. P. Distribution of Amino Acids in a Lipid Bilayer From Computer Simulations. Biophysical Journal 2008, 94, 3393–3404. (63) Pogorelov, T. V.; Vermaas, J. V.; Arcario, M. J.; Tajkhorshid, E. Partitioning of Amino Acids Into a Model Membrane: Capturing the Interface. The Journal of Physical Chemistry B 2014, 118, 1481–1492. (64) Vermaas, J. V.; Pogorelov, T. V.; Tajkhorshid, E. Extension of the Highly Mobile Membrane Mimetic to Transmembrane Systems Through Customized in Silico Solvents. The Journal of Physical Chemistry B 2017, 121, 3764–3776. (65) Kone, A.; Kofke, D. A. Selection of Temperature Intervals for Parallel-Tempering Simulations. Journal of Chemical Physics 2005, 122, 206101. (66) Van Der Walt, S.; Colbert, S. C.; Varoquaux, G. The NumPy Array: A Structure for Efficient Numerical Computation. Computing in Science and Engineering 2011, 13, 22–30. (67) Hunter, J. D. Matplotlib: A 2D Graphics Environment. Computing in Science and Engineering 2007, 9, 90–95. (68) Ferguson, A. L. BayesWHAM: A Bayesian Approach for Free Energy Estimation, Reweighting, and Uncertainty Quantification in the Weighted Histogram Analysis Method. Journal of Computational Chemistry 2017, 38, 1583–1605. (69) Habeck, M. Bayesian Estimation of Free Energies From Equilibrium Simulations. Physical Review Letters 2012, 109, 100601. (70) Hummer, G. Position-Dependent Diffusion Coefficients and Free Energies From 42 Pursuant to the DOE Public Access Plan, this document represents the authors' peer-reviewed, accepted manuscript. The published version of the article is available from the relevant publisher. Bayesian Analysis of Equilibrium and Replica Molecular Dynamics Simulations. New Journal of Physics 2005, 7, 34. (71) Price, W. S.; S¨oderman, O. Self-Diffusion Coefficients of Some Hydrocarbons in Water: Measurements and Scaling Relations. The Journal of Physical Chemistry A 2000, 104, 5892–5894. (72) Bonoli, L.; Witherspoon, P. A. Diffusion of Aromatic and Cycloparaffin Hydrocarbons in Water From 2 to 60.deg. The Journal of Physical Chemistry 1968, 72, 2532–2534. (73) Griepernau, B.; Leis, S.; Schneider, M. F.; Sikor, M.; Steppich, D.; B¨ockmann, R. A. 1-Alkanols and Membranes: A Story of Attraction. Biochimica Et Biophysica Acta (BBA) - Biomembranes 2007, 1768, 2899–2913. (74) Kondela, T.; Gallov´a, J.; Hauß, T.; Barnoud, J.; Marrink, S.-J.; Kuˇcerka, N. Alcohol Interactions With Lipid Bilayers. Molecules 2017, 22, 2078. (75) Odinokov, A.; Ostroumov, D. Structural Degradation and Swelling of Lipid Bilayer Under the Action of Benzene. The Journal of Physical Chemistry B 2015, 119, 15006– 15013. (76) Re, S.; Nishima, W.; Tahara, T.; Sugita, Y. Mosaic of Water Orientation Structures at a Neutral Zwitterionic Lipid/Water Interface Revealed by Molecular Dynamics Simulations. The Journal of Physical Chemistry Letters 2014, 5, 4343–4348. (77) Orsi, M.; Sanderson, W. E.; Essex, J. W. Permeability of Small Molecules Through a Lipid Bilayer: A Multiscale Simulation Study. Journal of Physical Chemistry B 2009, 113, 12019–12029. (78) Shinoda, W.; Mikami, M.; Baba, T.; Hato, M. Molecular Dynamics Study on the Effects of Chain Branching on the Physical Properties of Lipid Bilayers: 2. Permeability. Journal of Physical Chemistry B 2004, 108, 9346–9356. 43 Pursuant to the DOE Public Access Plan, this document represents the authors' peer-reviewed, accepted manuscript. The published version of the article is available from the relevant publisher. (79) Chipot, C.; Comer, J. Subdiffusion in Membrane Permeation of Small Molecules. Scientific Reports 2016, 6, 35913. (80) Zeppieri, S.; Rodr´ıguez, J.; L´opez de Ramos, A. L. Interfacial Tension of Alkane + Water Systems. Journal of Chemical & Engineering Data 2001, 46, 1086–1088. (81) Petelska, A. D. Interfacial Tension of Bilayer Lipid Membranes. Central European Journal of Chemistry 2012, 10, 16–26. (82) Ma, Y.; Ghosh, S. K.; DiLena, D. A.; Bera, S.; Lurio, L. B.; Parikh, A. N.; Sinha, S. K. Cholesterol Partition and Condensing Effect in Phase-Separated Ternary Mixture Lipid Multilayers. Biophysical Journal 2016, 110, 1355–1366. (83) Uline, M. J.; Longo, G. S.; Schick, M.; Szleifer, I. Calculating Partition Coefficients of Chain Anchors in Liquid-Ordered and Liquid-Disordered Phases. Biophysical Journal 2010, 98, 1883–1892. (84) Baumgart, T.; Hunt, G.; Farkas, E. R.; Webb, W. W.; Feigenson, G. W. Fluorescence Probe Partitioning Between Lo/Ld Phases in Lipid Membranes. Biochimica Et Biophysica Acta (BBA) - Biomembranes 2007, 1768, 2182–2194. (85) Mi, J.; Becher, D.; Lubuta, P.; Dany, S.; Tusch, K.; Schewe, H.; Buchhaupt, M.; Schrader, J. De Novo Production of the Monoterpenoid Geranic Acid by Metabolically Engineered Pseudomonas Putida. Microbial Cell Factories 2014, 13, 170. (86) Towns, J.; Cockerill, T.; Dahan, M.; Foster, I.; Gaither, K.; Grimshaw, A.; Hazlewood, V.; Lathrop, S.; Lifka, D.; Peterson, G. D. et al. XSEDE: Accelerating Scientific Discovery. Computing in Science & Engineering 2014, 16, 62–74. 44 Pursuant to the DOE Public Access Plan, this document represents the authors' peer-reviewed, accepted manuscript. The published version of the article is available from the relevant publisher. TOC Graphic 45 Pursuant to the DOE Public Access Plan, this document represents the authors' peer-reviewed, accepted manuscript. The published version of the article is available from the relevant publisher.
- the disintegration of a cell by rupture of the cell wall or membrane.
Lysis
Description
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7120914/
No comments:
Post a Comment
Note: Only a member of this blog may post a comment.