Fat bodies from fed or starved 4 hr flies were dissected in 1x PBS and then incubated in 100µL Lysotracker red DND-99 and DAPI for 20 minutes. A longer fixation was not used as paraformaldehyde could compromise the Lysotracker staining. Samples were mounted and immediately imaged using LAS AF software on a Leica SP5
confocal microscope.
8.
Brain paraffin sectioningAt least ten flies aged 7 days old from each genotype were anesthetized using CO2, after which a pair of dissection scissors were used to remove the head of each fly. Samples were then transferred to eppendorfs. 1 mL bouin solution (Polysciences, Inc.) was added to the samples as a fix buffer, following which the samples were rotated on shaker at room temperature for five to seven days. Kim whips were torn into small pieces and put into the eppendorf to soak samples.
Following this, the bouin solution was replaced with a leaching buffer (1M Tris (pH8.0), NaCl, ddH2O) overnight. Samples were then transferred into 70% ethanol and sent to the core pathology lab at the NHRI for dehydration, clearing, infiltration, embedding, sectioning and H&E staining. Six to seven serial sections of each brain were taken. The number of vacuoles in each section were counted, with the median
number of lesions calculated per brain.
9.
Electroretinogram (ERG)0, 7 and 14 day old flies were fixed in one direction on glass slides using non-toxic glue.
2M NaCl (for use as a conductive medium) was used to fill both recording and reference
electrodes. A reference electrode was placed in the torso of each fly whist recording electrodes were put over the retina.
Electrode voltage was amplified using a Digidata 1440A, filtered through a Warner IE-210, and the Clampex 10.1 software (Axon Instruments) was used for all recordings.
A light stimulus was provided in 1 second pulses via a computer-controlled red LED system (Schott MC1500). All experiments were conducted in duplicate or triplicate with at least 10 recordings completed for each genotype and experimental condition.
10.
Larval locomotion analysisFor each experiment, six L3 wandering larvae were selected and placed upon a 200 mm x 115 mm x 30 mm agar plate using a paintbrush at 25℃. Each agar plate contained a
solution consisting of 1% agar, 0.1M sucrose and brilliant blue dye (in order to provide a dark background for contrast enhancement). The plate was then transferred to a light, temperature and humidity controlled incubator. A camera was placed onto a tripod and
focused on the plate. Larvae were then allowed to move for four minutes, all of which was filmed. All genotypes were recorded with approximately the same circadian time period (between six and nine hours after lights on), although no circadian locomotive rhythms have been reported in larvae57.
Following video collection, the middle 2 minutes of each video was analyzed using
the wrMTrck plug-in for ImageJ58. At least three biological repeats were used for each genotype.
11.
RNA extraction and real-time PCR (qPCR) 1. RNA extraction.RNA was extracted from five L3 larvae per genotype and then homogenized in 500μl TRI-reagent (Invitrogen). 100μl CHCl3 was added and the samples were vortexed for 15 seconds, after which they were allowed to stand for up to 15 minutes at room temperature. The samples were then centrifuged at 12,000g for 15 minutes at 4 ℃.
The aqueous phase of the samples was transferred to a fresh tube and 250μl of cool 15 isopropanol was added. The samples were then allowed to stand for 5-10 minutes on ice before being centrifuged at 12,000g for 10 minutes at 4℃.
The supernatant was removed and the RNA pellet was washed in 250μl of 75%
ethanol and then dried for at least five minutes. 30μl of RNA-free water solution was
added before total RNA was quantified using a NanoDrop ND-1000 (Thermo Fisher Scientific Inc).
2. Reverse transcription -PCR (RNA→cDNA).
Following RNA quantification samples were transferred to new PCR tubes, each of
which contained 3μg of total RNA and 8.8 μl dH2O. The samples were mixed with 1μl 10X buffer, 0.1μl RNase inhibitor and 0.1μl DNase I and then incubated at room temperature for 15 minutes before 1μl of EDTA was added to terminate the reaction.
Samples were incubated at 65℃ for 10 minutes to inactivate the DNase I before 11μl
of the RNA sample was mixed with 0.1 oligo (dT) 12-18 primer and 1.5μl 10mM dNTP.
The mixtures were then transferred to a PCR machine at 65℃ for 5 minutes and
then put on ice in order to denature the RNA. 6μl 5X First-Strand Buffer, 3μl 0.1M DTT, 0.1μl RNase inhibitor, 0.2μl SuperScriptTMIII RT and 8.1μl dH2O were then added to each sample. Samples were kept in a PCR machine at 50℃ for 60 minutes, then 70℃
for 15 minutes to inactive the enzymes. The resulting product was stored at -20℃.
3. SYBR-system.
Samples were mixed with 3μl 10ng/μl cDNA, 0.5μl 5μM 5'-primer, 0.5μl 5μM 3'primer, KAPA 2x SYBR®Green dye master mix (a fluorescent double-stranded DNA -binding
dye used to track the progress of DNA amplification in real-time PCR experiments) and ddH2O. Samples were then loaded into 384 wells and an AB ViiA-7 Real-Time PCR system was used to detect Ct values. Data was analyzed using QuantStudio software.
Primer sequences for individual genes are shown in (Table 1.).
12.
Statistical analysisStudents’ t tests, chi-squared tests and Mann-Whitney tests were used for statistical
analysis. In general, p values <0.05 were considered significant whilst p <0.1 were considered as indicating a trend.
For the eclosion rate experiments, post-hoc pairwise chi-squared tests utilizing a Bonferroni correction were computed in R to compare WT eclosion rates with individual mutant eclosion rates59. All tests used a (corrected) significance level of p < 0.01.
Results
1. Generation of BCKDH knock-out mutants via CRISPR-mediated
mutagenesis
We generated BCAA-associated gene knockout flies using the CRISPR/Cas9 system.
Each of the BCKDH complex genes was targeted via CRISPR/Cas9-mediated genome editing by homology-dependent repair (HDR), which used two guide RNAs and a dsDNA plasmid donor. The excision began near the start codon and the
majority of the coding sequence was deleted by knocking in a cassette containing attPX, two STOP codons and 3xP3-RFP. As a selection marker, 3xP3-RFP was used to facilitate genetic screening. Figure 2 shows excision sites and the insertion cassette. The balanced stocks were labelled CG1673△/FM7, CG8199△/TM3, CG17691△/Cyo, CG5599△/FM7 and CG7430△/TM6B.
2. Mutant larvae display elevated levels of BCAAs (valine, leucine and
isoleucine).
To demonstrate that gene targeting was accurate, we investigated BCAAs level in mutant larvae. We found significant differences in the level of BCAAs in mutants as compared to wild-type controls (Figure 3). CG1673, CG5599, CG7430, CG8199 and CG17691 mutant lines showed 5.6~8.2 fold increases in leucine, 3~4.5 fold increases
in isoleucine and 2.3~4.3-fold increases in valine compared to wild-type (WT) flies.
These data validated BCAAs exactly accumulation in knock out larvae.
3. Loss of several BCKDH genes resulted in larval arrested development.
We found significant differences between controls and several mutant lines in terms of their eclosion rates (Figure 4). The majority of CG1673 and CG5599 mutant larvae could not eclose, with eclosion rates of just 7.3% and 7.6%. Those that did eclose died within a few hours. Furthermore, CG7430 mutant larvae were homozygous lethal, with no pupae found to have eclosed. However, CG8199 and CG17691 mutants were both found to be homozygous viable, with eclosion ratea of 82% and 96% respectively.
In total, eclosion rates were found to be significantly reduced in all mutant lines compared to WT (Figure 4). This implies that BCAA accumulation is toxic to immature forms of D. melanogaster, a similar phenotype to that observed in humans with the classic form of MSUD.
4. Vacuolar lesions increased in number in mutant brains.
As homozygous negative CG1673, CG5599 and CG7430 mutant lines were not viable to adulthood, we used heterozygous mutants for brain sectioning. We found that control
flies showed no lesions in any sample, whilst CG1673, CG8199, CG17691, CG5599, CG7430 mutants had median counts of 12, 16, 1, 23 and 16 respectively. We found a
significant difference in the number of lesions in CG1673, CG8199, CG5599 and CG7430 mutants as compared to WT (Student’s t test; p<0.05) (Figure 5.)
5.
Knockdown of BCKDH genes causes damage to neurons and results in neurodegeneration.Using the GMR-gal4 binary system to drive UAS-BCKDH RNAi expression we were able to observe morphological changes in photoreceptors. Since photoreceptors are light sensitive, we used a constant intensity of light to accelerate the neuronal damage process.
We found no significant morphological differences in the CG1673 knockdown line as compared to the control, whilst both CG8199 and CG17691 knockdowns showed slight decreases in the number of rhabdomeres after 14 days constant light exposure (6 and 6.6 rhabdomeres per ommatidium respectively) (Figure 6A). We also found a significant reduction for CG5599 and CG7430 knockdowns lines in the absence of light stimulation after both 7 and 14 days. Following 7 days of exposure, CG5599 and CG7430 knockdowns had 5 and 5.3 rhabdomeres per ommatidium. After exposure for 14 days, CG5599 and CG7430 knockdowns were left with only 3 and 1.6 rhabdomeres
per ommatidium.
Likewise, we found a similar phenotype in the electroretinogram (ERG) data.
Results from both CG5599 and CG7430 knockdowns indicated a reduction in depolarization and a loss of on-transients, with the trend in agreement with their morphology. In CG5599 and CG7430 knockdowns, ERG showed significant decreases
in depolarization amplitude compared to WT (Student’s t test, p<0.001) after 7 and 14 days. These indicates that the knockdown of CG5599 and CG7430 genes results in age-dependent neurodegeneration. For all other knockdown lines and conditions, we found no significant differences from control results.
6. BCKDH complex expression site in WT flies.
We used qPCR to investigate the distribution of BCKDH genes in fly heads, thoraxes and fat bodies. Three of the five genes (CG8199, CG5599 and CG7430) were highly expressed in muscle tissue (0.182, 0.183 and 0.695 fold) whilst the remaining two, CG1673 and CG17691,were more highly expressed in the fat body (0.05 and 0.39 fold) (Figure 8A). Furthermore, we found that irrespective of the tissue type, CG7430 was more highly expressed than all other genes (Figure 8B).
7. Lipid droplets accumulate in larval fat bodies.
dynamics and lipid metabolism, we measured lipid droplet size in L3 wandering larvae fat bodies. We found that lipid droplet size was significantly enlarged in CG1673, CG5599 and CG7430 mutant flies compared to controls (Figure 9.). Average wild type
lipid droplet area was found to be 123 μm2, whilst CG1673, CG5599 and CG7430 mutants had averages of 158, 202 and 184 μm2 respectively (Mann-Whitney test, p <
0.001). Other mutants showed no significant difference from WT flies.
8. BCKDH mutant found increased lipid droplets size but not CG1673
mutant in high protein diet.
To test whether changes in lipid droplet size occurred in response to a high protein diet, we dissected L3 larvae which had been provided with high protein food. High protein food was comprised of excessive soy protein including most essential
proteins. To further investigate lipid droplet size under different nutrient regimes, we fed larvae high protein food for 5 days after egg laying. BODIPY was used for staining. Lipid droplets were found to be bigger in all lines (except CG1673 mutants) having fed on the high protein food as compared to normal food.
9. Production of mmBCFA increased in MSUD mutants.
mmBCFA is derived from BCAAs and is involved in BCKDH activity. In C. elegans,
the critical mmBCFAs are C15ISO and C17ISO (Figure 10A.). To determine whether the lack of BCKDH activity resulted in mmBCFA deficiency, we measured the level of mmBCFA in L3 larvae by GC-MS. In contrast to C.elegans we found that mmBCFA-C17ISO was significantly increased in CG1673 mutants (Mann-Whitney test, p <0.05) (Figure10), suggesting that there is another enzyme responsible for mmBCFA synthesis
in Drosophila. However, we found no significant differences between WT and CG5599 and CG8199 mutant larvae – in these cases we were unable to detect C15ISO in our samples.
10. Starvation-induced autophagy may be downregulated in the larval
fat body.
In addition to its role in nutrient storage (which includes proteins, lipids and carbohydrates), the larval fat body is also highly active in autophagy. In late larval stages, cytoplasmic organelles are enclosed by a double-membrane vesicle and are degraded by autophagy to provide nutrients for metamorphosis. Within three hours of the onset of starvation however, a striking increase in both size and number of
vesicles has been reported. We therefore hypothesized that elevation of the level of BCAAs should inhibit autophagy.
To determine this, UAS mch-Atg8a was driven in the fly fat body by r4 gal4 in the knockout background. After four hours of starvation, puncta could be seen and calculated in individual cells. Mutants showed reduced Atg8a signals (Figure11.) with this result support by Lysotracker data.
In the well-fed diet condition, we could not see abundant autolysosomes (Figure
12A.) We found clear autophagy puncta in well-fed WT larvae, whilst mutant larvae exhibited significantly fewer puncta (Mann-Whitney test; p<0.05).
As expression level was not sufficiently high in controls, we therefore
investigated starvation-induced autophagy. Abundant and intense puncta appeared in all cells of WT flies, from about 2.2 puncta per cell to 8 puncta per cell after
starvation (quantification data shown in Figure 12B.). In contrast, mutant fat bodies had fewer puncta compared to controls with a range of 0.5- 4.7 puncta per cell, significantly less than in WT (Mann-Whitney test.; p<0.001).
Our results imply that elevated BCAAs could modulate activity of autophagy.
11. Larval movement capability was significantly reduced in BCKDH
mutants.
Muscle is the major position to carry out BCAAs metabolism. In human, infants who suffered from MSUD also had obvious muscle atrophy symptoms. In addition to muscle problem, neuron defects would also affect locomotor behavior.
BCKDH mutant larvae showed significant reductions in travel length and average
speed (Student’s t test; p < 0.001.). We also observed differences in crawling patterns (Figure13A.) compared to WT larvae, with mutant larvae tending to have
uncoordinated paths and fewer turns.
12. Mitochondrial morphology changes in muscle and fat body
The BCAA degradation pathway eventually produces acetyl-coA, which is sent to mitochondria to produce energy. Dysfunction of mitochondria may cause many problems such as developmental arrest, muscle atrophy and neurodegenerative diseases. To understand mitochondrial distribution, shape, and dynamics in muscle and in fat bodies, we used immunostaining and examined mitochondrial patterns.
In muscle, mitochondrial shape obviously different in all BCKDH mutants except for the CG8199 and CG17691 line. Mitochondria became smaller than wildtype in CG1673 mutant and CG5599 mutant showed diffused, dispersed phenotype. Whereas, CG7430 mutant appeared cluster mitochondria (Figure14.).
Mitochondrial mass changes were observed in the fat body. All of mutants had lower mitochondria mass than WT (Mann-Whitney test; p<0.001) (Figure15.). In addition, we found that CG5599 and CG7430 mutants had particular clusters of mitochondria (arrow in Figure15 E,F.).
13. Metformin treatment reduces the accumulation of leucine and isoleucine in E2 mutant larvae.
To determine the efficiency of Metformin in Drosophila, we measured BCAA levels (using LC-MS) in L3 WT and CG5599 mutant larvae 24 hours after Metformin (10mM) treatment. Metformin effectively rescued accumulation levels of leucine and isoleucine in E2 mutants, with no significant differences being found between
controls and mutants that had been treated with Metformin (Mann-Whitney test;
p<0.001). On the other hand, valine levels were found to increase in mutants after Metformin treatment (Figure 16.).
14. Metformin treatment rescues locomotor behavior in mutant larvae.
Larval locomotor defects were previously found in most mutant larvae lines (Section 11). We thus tested if Metformin was also able to rescue this phenotype. After 3 days of 10mM Metformin treatment, CG5599 and CG7430 mutant larvae demonstrated
more straight locomotor behavior than untreated ones (Figure 17A.). Larval travelled length was also rescued in mutant lines treated with Metformin but not average larval speed, suggesting that Metformin could recover mutants discontinuous mobility.
(Student’s t test; p < 0.001.) (Figure 17B.).
Discussion
Here, we generated a novel model of MSUD in D. melanogaster, then validated it by comparing molecular, electrophysiological and behavioral phenotypes of mutant flies with other animal models/ humans.
Knock out flies for the CG1673 (BCAT), CG5599 (E2 component) and CG7430
(E3 component) genes showed severe developmental problems. Homozygous flies were unable to eclose from pupae in knockout lines under normal feeding conditions.
This is similar in some respects to the human form of MSUD, with consumption of protein potentially leading to lethality in infants11. On the other hand, CG8199 (E1α component) and CG17691 (E1β component) mutants were homozygous viable at all developmental stages. Both mutants were E1 component knockouts however; as both subunits serve the same function, it is potentially unsurprising that inactivation of only one of these genes has little or no effect on the biological phenotype.
MSUD causes not only metabolic problems but also central nervous system disorders. In humans, MSUD patients have been shown to suffer from both white matter and neuronal injuries in magnetic resonance imaging studies; this includes extensive brain edema and pathological changes in the basal ganglia60,61. Neuronal defects are therefore a significant risk for MSUD patients, with even the most effective available treatment (liver transplantation) only proven effective for preventing acute crises and
unable to reverse psychiatric disease62. Similar to this phenotype, we found the appearance of vacuoles in brain paraffin sections of mutant flies. Furthermore, eye morphology changes revealed severe injury traits in mutant larvae, particularly CG5599 and CG7430 mutants. This phenotype became more serious when intense light was supplied for 7 and 14 days, suggesting the absence of these two subunits led to
neurodegeneration.
Additionally, analysis of our ERG data on photoreceptor function lent credence to the morphological data. Reduction of ERG depolarization amplitudes and absence of on- and off- transients resulted in neuronal defects. These data indicate that CG5599 and CG7430 could play a critical role in the normal function of neurons.
High BCAA levels are often found in obese, insulin-resistant states and in Type 2 Diabetes patients35. BCAAs thus became as a biomarker to detect type 2 diabetes.
However, the mechanism of how BCAAs elevated in obesity individual is still unclear.
There is a hypothesis to explain increased plasma BCAA levels to insulin resistance linked to activation of mTORC1: excessive nutrients lead to increase plasma level of leucine which together with insulin activate mTORC1 and S6K1. Persistent activation leads to serine phosphorylation of IRS-1 and IRS-2, which interferes with signalling and might target IRS1 for proteolysis via a proteasomal pathway. The resulting insulin resistance increases demand on insulin to expenditure overmuch glucose. Long-term
demand for insulin secretion, ultimately fail to produce sufficient quantities of insulin and lead to the T2DM18(Appendix v.).
In adipose tissue, obesity was associated with declines in BCATm (homologous to CG1673) and BCKD E1α (homologous to CG8199) protein concentrations in
ob/ob mice and Zucker fatty rats37,63. However, we found BCAT (CG1673) and two
BCKDH (CG5599 and CG7430) mutants were associated with increasing lipid droplet size in fat body which is a sign od obesity. Due to the absence of CG1673, CG5599 and CG7430 in adipocytes, we hypothesized that elevated BCAAs activated dTOR, which
downregulated autophagy and protein/lipid synthesis. Therefore, decline of protein and lipid turnover rate made excessive lipid storage in fat body. We would also like to investigate whether elevated BCAAs regulate insulin resistant and hyperglycemia in the MSUD Drosophila model in the future.
In order to investigate autophagy in MSUD mutants, we used an Atg8a marker and Lysotracker to monitor the progress of autophagy and formation of lysosomes. Here, we found that Atg8a marker expression was markedly decreased in mutant lines, suggesting that autophagy was downregulated. We hypothesize that activation of dTOR resulted in an inhibition of autophagy. Misregulation of autophagy possibly represents protein degradation imbalances in the fat body and mitochondria homeostasis in muscle.
Alteration of muscular, neural, and olfactory systems should affect locomotor
behavior. Analysis of mutant mobility could therefore help to increase our
behavior. Analysis of mutant mobility could therefore help to increase our