Update december 2017

Unfortunately, I haven’t been spending much time in the lab and was out giving talks and writing pieces for the newspaper. These weeks, I will have more time for science and I want to focus my attention on identifying any glycogen-degrading enzymes expressed by L. crispatus.

1 Screen of 20+ L. crispatus isolates for glycogen consumption.

In a project by Remco Kort together with TNO, VU University Amsterdam and GGD Amsterdam, 20+ L. crispatus strains were isolated from women with or without BV (bacterial vaginosis). The goal was to find metabolic and genetic characteristics that could predict whether a L. crispatus was more likely to be found in a BV background or in a Lactobacillus dominated background. Its genomes were sequenced and student Jorne Swanenburg, supervised by Remco Kort and Douwe Molenaar, has performed the assembly, annotation, quality control and analysis. The sequences are expected online at GenBank anytime soon, he is currently finalizing his thesis.

In addition, Charlotte van der Veer (PhD student at the GGD) has performed a preliminary API test to analyze the metabolic signatures of the strains. The API test revealed that there is quite some difference between the ability of L. crispatus strains to break down glycogen. This is relevant since glycogen is abundant in the vagina of reproductive-age women with a Lactobacillus-dominated community. Glycogen could function as a carbon source for certain L. crispatus strains to produce lactic acid and acidify the environment.

I am currently running growth experiments with these strains to verify the results of the API test and find differences amongst the strains. Previously, I have found the DSM strain (20584) to be able to grow on glycogen and I also found starch degrading activity after growth on glycogen both in supernatants as well as in the pellet. After growth on glucose, this activity was absent.

2 Genome analysis

Next, I am looking for genetic differences amongst this group of strains that correspond with their (in)ability to utilize glycogen for growth. For now, I am focusing especially on this gene, annotated for now as a “type 1 pullulanase”, expected size ~140 kDa. Previously, this gene turned up in a very small comparison between four L. crispatus strains from the Human Microbiome Project, that I carried out at the Washington University St Louis. There are some interesting differences between the genomes of the strains around this gene. Some strains lack a copy of this gene, some have one or two copies, and some have a copy that lacks the first 8 amino acids at the N-terminus. By the way, this protein has a SLAP domain, which means that it might be associated with the S-layer. Although I know very little about S-layers, it does look like the S-layer proteins are the ones that at the outer most layer of the cell wall which (warning: speculation ahead!!!) may be associated with the activity being present in both the supernatant as well as the pellet.

3 Protein purification

I hope to start my first attempts towards protein purification before the new year. I will try to fractionate all proteins in the supernatant since I hope/expect to find the glycogen-degrading activity amongst the larger proteins. I first need to find out whether I can continue using the NYCIII medium for this. It contains horse serum so has many enzymes of <100kDa (albumin, globulins). In case I do find the activity amongst the larger proteins I won’t need to try any other media. First, I will try some bench-top size exclusion chromatography using cross-linked sepharose. Wish me luck!





Update september 2017 ongoing experiments

Time for an update on last month’s progress!

LAB symposium

I attended the LAB symposium in Egmond aan Zee. Very interesting talks on a broad range of topics, from bacteriophages, to bacteriocins, to host-microbiota, to flavor and texture of lactic acid bacteria in food applications. I especially enjoyed the talk by Jens Walter who gave a broad overview of the natural history and lifestyle of the genus Lactobacillus. He mentioned that Lactobacillus iners, with the smallest genome amngst the lactobacilli, was “on its way to become an obligate symbiont”, given its loss of genes and functions. In the paper that was connected to this talk it says that the lifestyle-associated traits L. iners has “Fe-S—defense against peroxide, glycogen fermentation, adhesion”. However, I know of no published evidence that L. iners is capable of glycogen fermentation. Perhaps it is anecdotal, or just based on the genome.

Unfortunately, there were very few presenters (both posters and talks) who gave attention to vaginal LAB. I am convinced that the enormous effect of lactobacilli on reproductive health should in the future be a bigger part in the symposium and I am determined to create a lot of insight in the mechanisms underlying this colonization.

The symposium came to an abrupt and tragic end on Thursday after it became clear that one of the Irish attendees had been involved in a tragic accident after a midnight swim. Everyone was shocked about this tremendous loss in the LAB community. His name was Alan Lucid, may his memory be a blessing to his family and loved ones.

Growth of vaginal lactobacilli on glycogen

I have been growing L. iners myself, and have some preliminary indication that L. iners can grow well on glycogen, like Jens Walter wrote in his review. I also have a third replicate for L. crispatus growth on glycogen/water/glucose showing that this strain DSM 20584 can increase its cell density (grow?) on glycogen just as well as it can on glucose. (I updated the previous blog with the third replicate).

Organic acid analysis HPLC

Below you see the chromatogram of the UV detector of a L. crispatus supernatant after 24 hour of growth on NYCIII medium supplemented with either 0.5% glycogen, or water and L. iners grown on glycogen, and medium with glycogen as a control. I wanted to make sure that glycogen is indeed converted to lactic acid, and that these strains main product is lactate. At the moment I have only performed this analysis on one single experiment. This chromatogram indicates that lactate is the dominant organic acid produced by L. crispatus and L. iners. An acetate peak (expected at RT 18.9) is absent and also on the RID detector there are no big peaks detectable apart from some glucose in the medium and some lactate. I have not been able to identify all different peaks in the chromatograms. Especially the peaks at RT 21,8 and 23,9 have my interest since these might show some organic acid being produced and consumed, respectively.

Below is an overview of lactate and glucose present in each supernatant. Beware, this is a single experiment, I will have to replicate these experiments!

Starch degradation

I continued using starch to detect any possible glycogen degradation activity in the supernatants. I believe I can conclude that supernatants of L. crispatus cultures that are grown on glycogen (and not on glucose or water) have starch degradation activity, but will only breakdown ~60* of a 1% starch solution.. Unfortunately, I do  not have calibration curves for all measurements dates, so I am just showing optical density of the starch solution for now.  

Preliminary results show that this activity is sensitive to a freeze/thaw cycle. It is also present in the pellet of L. crispatus and I have not been able to detect any degradation activity in L. iners supernatants (on glycogen/water/glucose). Lastly, in answer of the question posed in the previous blog (is amylase activity product inhibited): addition of glucose to the starch assay did not result in lower activity. These are only preliminary results and I am looking into these initial observations, I do not have sufficient replicates yet to be able to say anything with certainty.

Work discussion

On September 11th I gave a work discussion to the group here at the VU. Very excited to share the data from my postdoc in St Louis, and all new ideas. I also gave this presentation at a few different meetings of clinicians. I spoke at the regular morning gynecologists meeting at the academic hospital VUMC, where I was invited by dr. Nils Lambalk. I was also invited by dr Bing Thio and dr. Marinus van Praag to speak at the “Brugge Dagen”, the yearly meeting by Dermatologists of the ErasmusMC. Dermatologists in the Netherlands treat most vulvo-vaginal disorders. These talks for clinicians were in Dutch, I will upload an English version of this talk to Figshare at a later stage.

Thoughts about product inhibition of amylase.

From the two preliminary growth experiments it seems that the strain of L. crispatus I am using is able to breakdown glycogen and use it for growth. If this observation holds (I am currently repeating it a third time, running the starch assay again and running the supernatants on HPLC to measure lactate production), it poses lots of questions for what this means in practice. Although the results have not been robust, it looks like starch-degrading activity is only present when L. crispatus was grown on glycogen and not when the cultures were grown on glucose. This immediately makes me think of carbon catabolite repression: the mechanism where bacteria in the presence of glucose shut down the expression of enzymes metabolizing other, less preferential, carbon sources.
My colleague Jurgen Haanstra posed an alternative hypothesis in the hallway yesterday. This doesn’t necessarily have to be regulation at the transcriptional level. This can also just be caused by product inhibition on the enzyme level! (thanks Jurgen, I really appreciate it). A few minutes later he sent me this paper from 1986 (cause that’s how he rolls):  Glucose feedback inhibition of amylase activity in Aspergillus sp. and release of this inhibition when cocultured with Saccharomyces cerevisiae.

Here, they were unable to measure amylase activity in the supernatants of this fungus, but when they dialysed the enzymes (i.e. replacing the liquid, while retaining the enzyme) amylase activity was back. If this phenomenon is also going on with the glycogen-degrading enzymes of L. crispatus this could not only explain why I am not seeing activity in glucose grown wells, but also why in the starch assay untill now only about half of the starch was degraded. If glucose accumulates, enzyme activity will seize. Pretty basic biochemistry actually. (Hundred years ago someone could have been doing the experiments I’m doing now, nonetheless it is pretty exciting. And as far as I am aware, pretty novel too.)
So, I am first going to simply add some glucose to the starch assay and also other sugars such as galactose and maltose. Maltose is also a breakdown product of glycogen/starch metabolism, so perhaps this will also provide some inhibition. Will keep you updated!

First growth experiments with L. crispatus (warning, unreviewed/unvalidated data) UPDATED

UPDATE 25/9/2017 I have a third biological replicate for L. crispatus growth that confirmed previous findings. I added a graph with average +/- standard deviations.

So, I have been growing the isolates that I ordered (see previous post) from DSM with mixed success.

-The bad news is that the Gardnerella vaginalis -80°C glycerol stock seems to be in bad shape. I inoculated twice successfully from this stock on NYCIII growth media (both liquid and agar plates), but after this it took multiple days for the culture to grow and lately they haven’t grown at all. I suspect this has to do with the aerobic condition in which they are stored. I will retry with plates and media that I will preincubate in anaerobic (N2+CO2) conditions, but there will be some influx of oxygen while inoculating, and at this point I am not sure that it is feasible to handle an study G. vaginalis outside of an anaerobic chamber. Other methods are still optional, such as using closed infusion flasks.

-The good news is that both Lactobacillus crispatus and Lactobacillus iners stocks seem to be growing well. I am using NYCIII medium for both, Lactobacillus crispatus grows within ~16 hours anaerobically, while Lactobacillus iners requires ~48 hours.

I present my first experiences with growing L. crispatus on glycogen.

Some background

The goal of this endeavor is to test whether vaginal bacteria can grow on glycogen, and are able to convert glycogen to lactic acid. Glycogen is an important carbohydrate present in the vagina, shown by old [1] and new research [2]. It is not yet understood how lactobacilli acidify the vagina of reproductive age women, and glycogen metabolism could be an important mechanism. Previously, Spear et al [3] have shown the presence of glycogen degrading enzymes in the vagina. The lactobacilli that were tested in this study showed no glycogen degrading capacity . Spear et al, asserts that the host excretes amylases (glycogen degrading enzymes) in order to assist acidification by Lactobacillus. However, the researchers did not use the most often encountered vaginal lactobacilli: Lactobacillus iners and Lactobacillus crispatus. Here I show initial growth experiments of L. crispatus on glycogen and amylase assay of the supernatants. I have gathered some evidence that Lactobacillus crispatus DSM 20584 strain can utilize glycogen as a source for growth.

Methods and results

I have used NYCIII medium supplemented with either 0.5% glycogen or nothing (water, negative control) or 0.5% glucose as a positive control. To this end I have pipetted 100 uL of a 5% glucose solution or 5% glycogen solution or water and 900 uL of a 1.1x NYCIII medium where I have left out the glucose (see protocols). I have inoculated this with 100 uL of a preculture of L. crispatus DSM 20584 grown for >24 hours at 37°C anaerobically, without shaking. . For every condition, I used three wells as technical replicates. I also have empty controls without cells. Glycogen makes the solution a bit hazy so I wanted to make sure that any increase in optical density is not due to the glycogen itself but this was not the case. I have performed this experiment on two different occasions.
I mixed the culture by pipetting with a 1 mL pipet and diluted the culture 10x with water in a flatbottom 96-well plate to measure the cell density in the plate reader at OD 600. It looked like this

The data look like this: I’ll upload this to a platform once I have a bigger data file to share.

Updated graph with mean +/- standard deviations, unpaired t-test comparing OD600 on glycogen compared to water shows p-value of below .0005.

The optical density is increased when glycogen is added to the media, comparable to the increase by glucose. This is an indication that this particular strain can use glycogen for growth. Next, I wanted to see if I could detect the enzymes that L. crispatus uses to break down glycogen for uptake and metabolism.

To this end, I spun down the cells for 20 minutes at 4°C and maximum speed (4754 rcf, 4499 rpm) and transferred the supernatants to plates I stored at -20°C. I wanted to test whether L. crispatus excreted any glycogen degrading enzymes in the supernatant. I added 50 uL of supernatant to 150 uL of a 1% starch solution in “amylase buffer” (100 mM Na-acetate+5mM CaCl2, pH 5.5). Starch is not the same as glycogen, the polymer has a different structure. The backbone of starch consists of the polymers amylose and amylopectin, which also consist mostly of glucose units but which are branched differently. Looking at Google images for these three polymers gives a good idea

I hope and expect that the enzymes that degrade most of the bonds between glucose units of glycogen will breakdown the bonds in amylose and amylopectin too. Starch has the advantage that it can be easily detected using iodine, an experiment that many of us already did in elementary school. I incubated this for ~24 hours at 37°C. I also checked after one and two hours but did not observe any breakdown of starch. I didn’t use any amylase control yet, only a reference calibration curve for starch. After incubation I added 10 uL, to 290 uL of iodine working solution and measured the absorption at 600 nm. I made a calibration curve using dilutions of the 1% starch concentration.

This experiment is not robust at all, since three times I measured three different things. So I will not share protocol and data yet. During the first measurement this calibration curve was not completely straight, not sure why, but the controls (starch with the supernatants of the empty control) showed an average of 7.6 gram/L starch, which is the concentration you expect. (150 uL of 10 gr/L starch + 50 uL of starchless liquid). However, in the wells with the L. crispatus supernatants something curious is going on. It seems that the amylase activity is found in the wells grown on glycogen whereas the wells grown with no carbon source (water) show the same concentration of starch as the controls (no breakdown). The glucose wells show some increase in starch. I am not sure how this happened, perhaps it has to do with pH, lactate presence, or even some growth in the wells. The wells did not look turbid, but growth can not be entirely ruled out.

The second time I did not freeze/thaw the supernatants but used fresh after spinning down the culture. I saw starch breakdown in the glycogen wells, indicating that L. crispatus excretes some soluble glycogen degrading enzyme in the supernatants (amylase?) whereas the H2O and glucose well did not show signs of starch breakdown. A third time I performed this experiment again with frozen sups I did not see any breakdown of starch.

So, I have to optimize this method, run controls with diluted amylase. I have to find out what’s going on with the increase in OD and finetune the protocol to get reproducible results. Perhaps it will turn out that this method is not robust enough, and I will have to resort to another way of measuring amylase activity.

On my long grocery wishlist with experiments:

  • Improve amylase detection.
  • Perform a HPLC on the supernatants to determine metabolite concentrations and gather more evidence of glycogen metabolism.
  • Use HPLC to measure amylase products (glucose, maltose, trisaccharides etc).
    Continue with Lactobacillus iners and Gardnerella vaginalis to study its glycogen degrading abilities.
  • I would like to partially purify the amylases to perform proteomics, we have applied for a “Hotel”-grant at ZonMW to perform this analysis together with Winclove probiotics at Radboud UMC.
  • Lastly, we want to extend the analysis to other strains of crispatus that were isolated by Remco Kort and his student Jorne Swanenburg in order to find the gene and enzyme involved in this amylase activity. Happy to announce that the biomedical/filosophy student Noa Fuks has offered her help to track down the amylase, while teaching herself some bioinformatics. Good luck Noa!

So, it’s a long grocery wishlist of experiments. If you have any comments, suggestions for alternative amylase activity assay, ideas, theories, criticism, please drop me a line.

1. Cruickshank, R., The conversion of the glycogen of the vagina into lactic acid. Journal of Pathology and Bacteriology, 1934. 29.
2. Mirmonsef, P., et al., Free glycogen in vaginal fluids is associated with Lactobacillus colonization and low vaginal pH. PloS one, 2014. 9(7): p. e102467.
3. Spear, G.T., et al., Human alpha-amylase present in lower-genital-tract mucosal fluid processes glycogen to support vaginal colonization by Lactobacillus. The Journal of infectious diseases, 2014. 210(7): p. 1019-1028.

Ordering and stocking strains

Getting started! I have ordered, received and stocked 3 strains at the DSM strain library. Unfortunately they do not have as many different options for vaginal bacteria, compared to all the isolates from the Human Microbiome Project for instances, and the strains I ordered were not all isolated from the vagina. I hope that they will display the same characteristics and can be used as a reference and to assist in setting up my methods. In the future I hope to with more vaginal isolates. I have

To grow the strains I used MRS medium (liquid and plates) for L. crispatus and NYCIII medium for G. vaginalis and L. iners. Unfortunately I do not have access to an anaerobic chamber so I have been using an anaerobic jar. After closing the jar I use 3x: pulling a vacuum (minimally 0.8 bar) and filling with N2+CO2 gas. I grew plates and tubes over the weekend (72 hours) at 37 degrees and stocked them using 0.5 mL 60% glycerol + 1 mL of culture and stored them at -80 degrees. I am not sure how much the oxygen will affect the bacteria during handling and storage, especially Gardnerella is sensitive to oxygen. I will check later if the stocks lead to good growth. I will upload protocols for the media, the stocking and the anaerobic jar soon.

First steps

Hello everyone, first post of my “open kitchen science” research initiative. I am very excited to announce that I have started working at the Free University of Amsterdam, within the Host-Microbiome group of Prof. Remco Kort at the Microbiology department. I will keep a regular update on the progress here. Currently I am still finding my way around in the labs, looking for chemicals, getting the right strains and looking what equipment is available. I will also start filling the website with background information on metabolism of vaginal microbes and the question I want to answer with my experiments.