Biochemistry

Picture kindly supplied by Bruno Ribeiro

In lands afar where coffee beans grow,

Exotic flavours in each cup do show.

From the mountains high to the valleys deep,

A symphony of tastes for us to keep.

Let us savour these exotic brews,

Each sip a story, a vibrant muse.

In coffee's world, a taste so grand,

A symphony of flavours, at our command.

ChatGPT, 2024

The reservoir of aroma potential, discussed in the previous article, should be tapped, but when is the best time to do so? Unlike wine or fruit juice in which these compounds are typically studied, coffee is not made from the fruit's flesh but rather the seed, introducing the importance of timing in relation to coffee cherry development as well as potentially post-harvest processing. For instance if these compounds occur naturally in the seed, it is important to determine when they are formed during development. Whereas if they are generated from the mesocarp of the coffee cherry, they would have to diffuse into the seed during post-harvest processing before impacting coffee’s quality. In this case, their molecule mass serves as an indicator of their diffusion rate.

A recent study by CIRAD and its collaborators’ (2024) found that the monoterpene limonene in a specific Geisha variety is linked to the genetic expression [1] within the coffee seed. This study challenged my scepticism about the seed's role in terpene production, as existing literature strongly associates terpene synthesis [2] with the maternal tissue of the fruit.

Nevertheless, the connection of these pathways with the maternal tissue of the coffee cherry, continues to make more sense to me, than its direct production in the seed, as it further supports the compounds’ association with the fruit's self-defence mechanism. The defence mechanism functions in that the glycosylated bound volatiles (GBVs) are kept in a separate cellular compartment to that of the enzyme [3] capable of releasing their defensive character [4], i.e. the aromatic secondary metabolites we enjoy. To release the aroma, the compartments need to be ruptured, which typically happens when the fruit is attacked by a predator.

Expression of Key Genes Involved in Terpene Pathways in

Critical Tissues of Coffea arabica

modified from Silva et al., 2020

Based on the role of these compounds, I hypothesize that coffee cherries will not diverge greatly from the pattern observed in literature [5], where the expression of terpene’s biochemical pathways will predominately occur in the coffee cherry's mesocarp, or maternal tissue. In other words I suspect terpene synthesis within the seed is secondary to those occurring within the fruit’s flesh, making the latter potentially more influential to cup quality relative to the former. As shown in the figure above, Silva et al. (2020) recently provided evidence supporting this hypothesis by demonstrating that key genes within the two terpene pathways were preferentially expressed in the maternal tissue of the flower and perisperm of C. arabica [5.5]. The expression of these pathways within the seed may even be an evolutionary mutation or adaptation to the plant’s surroundings, possibly making them variety dependent as suggested by CIRAD and colleagues’ 2024 study. Naturally, further research is required to substantiate these suspicions and flesh this field of coffee research out.

[1] The gene encodes an enzyme associated with the biochemical pathway terpene synthesis.

[2] I am taking terpene synthesis as a representative example of the formation of secondary metabolites within coffee as they are linked to the composition of the cherry.

[3] Glucosidase. As there are also microbes capable of generating these enzymes, GBV’s within the coffee mesocarp may also be liberated through fermentation. Fermentation may also liberate GBV’s by lowering the pH resulting in the acid hydrolysis of the glycosidic bond.

[4] This is the volatile toxic compound discussed earlier.

[5] It is also entirely possible that the simultaneous expression of terpene synthesis pathways in both the maternal and offspring tissues have been under research, leading a potential bias in the literature introduced by the nature of the previous products.

[5.5] Terpenoid pathways also contribute to the production of phytohormones, so they are not solely responsible for producing phytochemicals. This means they may have other important functions as well.

Diffusion

Nevertheless, if this hypothesis is correct and these pathways predominately occur within the maternal tissue of the coffee’s mesocarp this would also mean that to benefit from this aroma pool these compounds would need to diffuse into the beans – before or during post-harvest processing. However, which compounds do we want to diffuse into the beans: the free volatiles, e.g. monoterpenes, or their glycosylated equivalent? Given that only compounds that remain after roasting can impact cup quality, let's explore what might happen to GBVs and their un-glycosylated counterparts during thermal processing.

Literature suggests (Haure et al., 2022) that GBVs decrease during thermal processing [6] liberating the bound aromas [7], however it remains unclear to what extent this occurs and whether it varies based on the roast degree [8]. Considering that thermal processing can degrade GBVs, it may be reasonable to assume that the free form of these compounds accumulate during roasting [9].

The unlikely loss of secondary metabolites during roasting, makes it seem reasonable to assume that to benefit from this reservoir of aroma within the maternal tissue, that either form needs to diffuse into the bean.

This conclusion would have several practical implications,

- Coffee cherries should not be injured [10], as this leads to the degradation of GBVs and loss of the volatiles. In other words, do not trigger the fruit’s defence mechanism, if you do not want to lose these aromas. Applying this knowledge to current practices one would hypothesize, assuming the coffee is immediately processed upon harvesting, that coffee dried on the plant is most likely to benefit from these aromas, followed by naturally dried coffees, possibly fermented, and then washed coffees.
- As GBVs are formed during ripening of the coffee cherry, it is unlikely that coffee cherries that are harvested prematurely will benefit from this aroma pool – these compounds have yet to be formed. Therefore the coffee cherries that have the greatest potential to benefit are those that have ripened completely.

GBVs Reported In Literature

You may have noticed that three of the GBVs share the same molecular weight. However, we can still differentiate between these molecules based on their retention time or their ability to move through the same media or environment. It should also be mentioned that a molecule of formic acid has been subtracted from each mass as they were reported as formic acid adducts (a common ionization aid).

- GBVs are large molecules, approaching the physiological diffusion limit of 500 Da (de Rosso et al., 2022), which means they will only gradually diffuse into the bean, relative to their un-glycosylated, smaller counterparts. This means that the longer the coffee cherry stays intact (mesocarp-endocarp contact time), the greater the chance is that more GBVs will have diffused into the beans. Consequently, the impact of GBVs on roast coffee quality would depend upon when the coffee cherries were harvested and on the specific variety [11]. Once the GBVs have diffused into the bean, they are unlikely to be recognized as food by the bean's own metabolism, meaning that this aroma pool will be preserved until roasting [12]. For further details on the bean’s metabolism during post-harvest processing, please see "Deconstructing Coffee Fermentation Part 3: Alternative Pathways." This suggests that the perceived quality associated with longer development times, particularly at higher altitudes and in agroforestry systems, may be partly due to the longer diffusion times these compounds require in order to travel into the bean.
- The bound volatiles within the green bean can be freed under relatively mild conditions (those used for juice pasteurization), which suggests that mildly roasting coffee (lightly roasting the coffee) should be sufficient to release this aroma pool within the bean. However, as their free form are exceptionally volatile, prolonging the roast will likely lead to their loss.

In summary, the implications of these potential outcomes all suggest that processing green coffee according to Yemeni's traditions, discussed in part 1, is most likely to lead to the accumulation of GBVs in the green bean and consequently enhance the coffee quality upon roasting. In other words the release of GBVs [13] upon roasting may be responsible for the exotic flavours (e.g. blueberry, blackcurrant) often associated with the elevated cup quality of the region of Yemen and Ethiopia.

[6] Add a few literature examples with the associated heat treatment

[7] The bound sugar molecules are also released and will likely degrade on their own or react further, depending on the roasting conditions, i.e. mild conditions may allow the sugars to participate in the Maillard reaction. This can be clarified through isotopically labelling the sugar moiety of GBVs and then impregnating green beans with these molecules. When this coffee is roasted and then compared to a control, where an unlabelled version was incorporated in the green bean in the same manner, one will observe shifts in the weights of the compounds incorporating fractions of the labelled molecule.

[8] I believe this is the case, which explains why the specialty coffee community prefers light roasts and why these unique aromas are seldom present in darker commercial roasts.

[9] The boiling points of monoterpenes are quite high. For instance, linalool has a boiling point between 194-197°C and geraniol between 229-230°C. While these compounds may not be lost during roasting, they may react further into disagreeable or sensory irrelevant compounds. Their fate remains uncertain, but if nothing happened dark roasts would still have a fruity character.

[10] This may also contribute to the seasonality of coffee quality as if excess rains during harvesting lead to the bursting of the coffee cherries, this would compromise the integrity of the coffee cherries affecting the aroma pool and thereby the coffee quality.

[11] A gap in knowledge exists regarding how coffee varieties differ regarding the diffusivity of the beans and parchment. Leading me to suspect that this may impact the rate of diffusion of these analytes as well as the time required.

[12] Naturally if there is an insect or fungal contamination, than these compounds can be degraded.

[13] . . . and retention if properly roasted. These compounds are volatile so they can always be lost if care isn’t taken.

Fermentation

One way of increasing the rate of diffusion is to reduce the size of the molecules, for instance by cleaving the glycosidic bond.

Fermentation can accomplish this in two ways:

1) Some microbes contain enzymes that can release the volatiles from their sugary bonds, and

2) Acidifying GBVs is a common analytical approach of liberating the volatiles for analysis, refer to the graph below. Therefore, it is possible that certain forms of fermentation that significantly lower the pH of the environment may also release these volatiles.

de Rosso et al., 2022

What remains uncertain is whether these compounds will leave the tank and partition into the surrounding air, or if whether they will preferentially diffuse/partition into the beans - this may be dependent on the bean's lipid content (terpenes are lipophilic, "fat loving"). In the absence of additional information it is difficult to hypothesize on the net effect of these conditions on coffee quality.

From a scientific perspective one would need to run a controlled set of experiments investigating,

The influence of the enzyme versus the acidification of the media, mimicking fermentation, on GBV release.
Sensory significance of any changes. Infusion of the modified media into green beans for sensory analysis.
Diffusion preferences of these compounds [14]. For instance using isotopically labelled linalool to understand the preferences of this class of compounds.

From a practical and safety standpoint [14.5], if you are fermenting in an open tank and you smell the aroma, this is not necessarily a good thing as it means that the aroma has escaped from the tank. It could be that there is sufficient aroma in the tank as well, but pay attention to the final product quality and if it is proportionate to the expectations you formed during processing.

What do you think? Please share your thoughts in the comments below.

[14] Refer to the fermentation section of this website.

[14.5] Chronic exposure to aroma compounds should be considered a workplace safety issue. There is a well conditioned linked to popcorn factories from chronic exposure to butter flavouring, diacetyl, which can be found in a variety of products, including coffee, resulted in scaring of the lung tissue, known as Bronchiolitis Obliterans, or popcorn lung, signigicantly affecting your breathing.

“Late-Harvest”

More recently Yemen’s traditional mode of processing has borrowed the term “late harvest” from the wine industry (Sanz-Uribe et al., 2017; Nespresso, 2012). Given the historical context, adoption of this term seems in bad taste as Yemen was and is an Islamic nation where alcohol is considered “haram” or prohibited by their faith. Seeing as this knowledge was most certainly lost as coffee cultivation extended beyond Yemen’s shores, it is entirely possible that the unfortunate association was simply an oversight by those reintroducing these practices into the coffee industry. Nevertheless, seeing as these practices do not originate from winemaking, but are ones that coffee can take full ownership of, I propose that the coffee community rebrand these practices, potentially to reflect its coffee and Yemeni heritage?

To understand which flavour profiles and aroma notes are associated with the coffees processed in this manner I used the term, “late harvest”. My literature search revealed that due to the lack of understanding of this process the valuable potential of GBVs are frequently lost or obfuscated by aroma attributes associated with further processing, making this process often indistinguishable from any other [15]. Essentially, this nullifies the added value applying these practices should have provided these coffees. In other words if you don’t know what a target looks like, how can you hit it?

Based on the product descriptions from commercial roasters [16], it seems that Nespresso's late harvest, high altitude, washed Colombian special edition coffee and Lavazza's Tales of Milano from "Central America" were the only ones that indicated an understanding of the expected chemistry involved in this process [17]. However, this doesn't mean that other roasters are oblivious to these processes or chemistries. For instance, Illy has not only been actively researching this subject for years (Bonnlaender et al., 2006; del Terra et al., 2013; de Rosso et al., 2022), but they have acknowledged the quality of these coffees on several occasions by recognizing them in the IllyCaffè awards. The award winning nature of these coffees can also be seen with their high rating at Cup of Excellence (COE) auctions.

[15] This confusion of flavour and aroma traits will make more sense once I have written the article series on the Maillard reaction.

[16] I must admit I only looked at the largest European roasters.

[17] Feel free to comment below if you have tried these products and find the product description inconsistent with the product itself.

What Can We Learn

If secondary metabolites serve as a source of aroma that can contribute to coffee quality, then considering their physiological location is crucial to ensuring their incorporation in the final product.

If these aromas are located outside the final product, the green bean, and rely on diffusion during ripening to enter, then the time required will depend on the size of the aroma-relevant molecule.

The question then becomes, "Which molecules hold the greatest potential of contributing to coffee's aroma quality after roasting, bound or or unbound volatiles?"

Discover The Latest Articles By Joining Dr. Kôfē's

Mailing List

Do You Have Any Questions?

Please feel free to reach out by email or . . .

(comments are not publicly posted)

Page updated

Google Sites

Report abuse